Research ArticleStrategic Information Sharing in a Dynamic Supply Chainwith a Carrier under Complex Uncertainty
Heng Du 1 and Ye Jiang 12
1School of Management and Engineering Nanjing University Nanjing 210093 China2School of Business Jiangsu Open University Nanjing 210011 China
Correspondence should be addressed to Heng Du 18795898791163com
Received 10 January 2019 Accepted 13 May 2019 Published 2 June 2019
Academic Editor Manuel De la Sen
Copyright copy 2019 Heng Du and Ye Jiang This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited
Whether to use an information sharing mechanism is investigated in a dynamic supply chain where onemanufacturer one carrierand one retailer are faced with uncertain yield demand and lead time during multiple periods Each member is modeled as anadaptive agent based on multiagent technique and their decisions can be adjusted timely to adapt to external environment Thereare two choices for the whole supply chain to deal with uncertain risks information sharing (IS) or no information sharing (NS)Under strategy 119868119878 the information aboutmarket demand and the retailerrsquos inventory can be sharedwithin the supply chain For eachstrategy the effects of yield demand and lead time uncertainties on costs of the supply chain and channel members are studiedIt is found that (i) it is rewarding for the upstream manufacturer to use a retailerrsquos shared information under uncertain yield ordemand (ii) however information sharing (IS) strategy sometimes should be abandoned for other members and the whole supplychain (iii) counterintuitively the increase of transportation time uncertainty benefits the retailer
1 Introduction
Information sharing is regarded as a prevalent businessstrategy to improve operations performance of the supplychain which has been successfully used in many industriesIt is widely acknowledged that information sharing cansimultaneously benefit the whole supply chain and eachmember [1] A classic case isNestle andTesco [2] Bymeans ofsharing sales data between partners Tesco sharply simplifiesthe organization procedures and Nestle also reduces theinventory cost Traditional information sharing focuses onthe relationship between sellers and buyers But with thedeepening of labor division more and more intermediatecarriers emerge and are authorized to delivery materialsinstead of upstream shipper firms The carrier a transportservice provider plays a significant role across the supplychain [3] For instance Fedex (a third party logistics inAmerica) collaborates with computer manufacturers (such asAppleDell and IBM) and retailers inTaiwan Some real-timeinformation is shared among them so that the profit of eachmember is raised [4] However it is not always the case for all
firms As an example of Yingte (a pharmaceutical companyin China) it ever failed to decrease cost by sharing infor-mation with a carrier and partners Consequently strategicinformation sharing should be used in the actual situationStrategic information sharing a flexible strategy is definedas two choices information sharing or not Namely it isnot always necessary to utilize information sharing strategysometimes sharing information should be adopted in thesupply chain but it should not be selected at other timesThus motivated by these practical observations it is one ofour goals to understand whether information in a multilevelsupply chain with a carrier should be shared
Information sharing decision is usually directly affectedby external complex uncertainty [5] Uncertain marketdemand and stock information are shared by many compa-nies to mitigate the bullwhip effect for instance PampG Wal-Mart and Cisco [6] To cater for time-sensitive consumersLampTT in Hong Kong presents own production data toupstream suppliers to eliminate the impact of uncertain lead-time risk [7] Nevertheless there are also some companies
HindawiDiscrete Dynamics in Nature and SocietyVolume 2019 Article ID 4695654 13 pageshttpsdoiorg10115520194695654
2 Discrete Dynamics in Nature and Society
not willing to provide information to others due to environ-mental uncertainties After all they are afraid that the simpleinformation sharing behavior may not cope with the complexuncertainties [5] Given the different attitudes toward theinformation sharing under an uncertain situation this paperexplores how multiple external uncertainties influence infor-mation sharing strategy for a supply chain such as uncertaindemand supply and lead time
The actual supply chain is proved to be a dynamicsystem [8] The complex dynamics are mainly reflected intwo aspects On the one hand the supply chain exists in anextremely uncertain environment where almost all externalelements vary all the time On the other hand the membersin the supply chain are autonomous individuals Sometimesthey adjust own decisions to adapt the dynamic exteriorcircumstance We refer to this supply chain as a dynamicsupply chain To the best of our knowledge the static optimaldecision is mainly concentrated on in conventional supplychain studies which are difficult to reflect the dynamic char-acteristic of the supply chain Hence unlike most researchesthe supply chain dynamics in a changing market that consistsof multiple periods however is the focus of this paper
An information sharing strategy is investigated in a three-level supply chain where one manufacturer one carrier andone retailer are faced with uncertain demand yield and leadtime There are two choices for the supply chain to manageexternal uncertainties information sharing (denoted by IS)or not (denoted by NS) Based on the method of multiagentmodelling we compare the total cost of the supply chainand each member in two cases to address the followingissues (1) Can the information sharing be beneficial tothe whole supply chain and each member in a dynamicenvironment simultaneously (2) What are the impacts ofsome uncertain risk factors on the information sharingdecisions We obtain meaningful management implicationsFor instance it is rewarding for the upstream manufacturerto use the retailerrsquos shared information under uncertain yieldor demand whereas information sharing strategy may beabandoned for the whole supply chain and other channelmembers and the retailer can obtain more benefits with theincrease of transportation time uncertainty
2 Literature Review
This paper is related to the information sharing in the supplychain and multiagent modeling
From the perspective of participants in the supply chainthe information sharing literature can be classified into twostreams The partnership between sellers and buyers is widelydiscussed in the first stream For example Cachon and Fisher[9] Lee et al [6] and Teunter et al [10] studied the valueof information sharing between upstream and downstreamfirms It is concluded that sharing information is beneficial toboth parties only under certain conditions Dejonckheere etal [11] Chatfield et al [12] Ma and Ma [13] and Zhao et al[14] analyzed the impacts of many factors on the well-knownbullwhip effect Aviv [15] Fildes and Kingsman [16] Traperoet al [17] and Sanders andWan [18] focused on how forecasterrors affect the information sharing whereas our paper is
different from these researches in that a carrier is regarded asa key supply chainmember here Particularly we consider theimpact of a carrier on the information sharing in a multilevelsupply chain The second stream is about the collaborationwith the third party logistics The typical studies are fewFor instance Wen [19] explored how to forecast shipmentof the carrier based on shared information However ourwork places emphasis on identifying complex uncertainfactors influencing the information sharing strategy Tyan etal [4] and Wen [3] qualitatively described the frameworkand competitive advantage of collaborative transportationmanagement (CTM) To be different quantitative study oninformation sharing behavior for a supply chainwith a carrieris conducted in our paper Chan and Zhang [20] and Li andChan [21] investigated the benefits of CTM in the mode ofmake-to-order (MTO) which is somewhat similar to oursIt is found that CTM lowers the total cost and risk for thewhole supply chain Yet the cost of each member is notdiscussed As a matter of fact it is necessary to guaranteeeach memberrsquos benefit in the case of information sharing Incontrast with [20 21] our differences are mainly displayedin four aspects (1) not only total cost of the whole supplychain but also that of each individual is examined as well (2)the manufacturer exogenous in their researches is served asan adaptive agent here which is more in line with practicalcases (3) the supply chain here is a make-to-stock (MTS)production system rather than the MTO system namely themanufacturerrsquos order is based on forecast (4) the effects ofmultiple uncertain risks are taken into account
Multiagent modeling is also correlated with our workThe traditional approaches about operations optimizationare widely adopted in the supply chain management whichattempts seeking the optimal decision Instead this staticoptimal behavior is usually not in line with practical casesAfter all the supply chain is a complex adaptive systemwhere each member has to be confronted with an uncertainsituation In addition practical members across the supplychain are bounded rational [22] who are hard to acquirecomplete information and find the best decision due tothe own ability In most cases adaptive learning throughpast experience is the common method to make decisionsMultiagent modeling (MAM) is a powerful and populartool to solve the complex dynamic problem owing to thedistinct strengths [23] Consequently MAM is introducedto depict the dynamic and autonomous features that weprimarily focus onThere have been representative literatureson MAM For example Swaminathan [23] Long [24] andYu and Wong [25] construct a framework to explore thesupply chain network dynamics Dogan and Guner [26] andHe et al [27] discuss pricing and ordering policies underdemand uncertainty In addition some other problems suchas inventory strategies [28] products management [29] andscheduling [30 31] are examined by many scholars as well
To sum up this paper contributes to the literature inseveral aspects First unlike many literatures the carrier isconsidered as a crucial member in a supply chain The issueof whether to share information with an intermediary carrierin a supply chain is investigated Each partyrsquos cost especially isstudied in detail Second we further explore the motivation
Discrete Dynamics in Nature and Society 3
ManufacturerRetailer
Information sharing platform
material flow
Carrier
material flow
information flow information flow
information flow
information flow information flow
ManufacturerRetailer
material flow
Carrier
material flow
information flow information flow
Order forecast Capacity forecastOrder pointadjustment
Strategy NS
Market demand and the inventory
Market
Market
material flow
material flow
Uncertain yield
Uncertain yield
Uncertain transportation Uncertain demand
Uncertain transportation Uncertain demand
Strategy IS
Figure 1 Two strategies IS and NS
to share information under external uncertainties To be spe-cific the impacts of uncertain demand yield and lead timeon information sharing are discussed Lastly the complexsupply chainrsquos dynamic and adaptive natures are captured inthis paper In particular each member is capable of alteringown decisions in a dynamic environment
3 The Model
31 The Overall Structure and Problem Description Considera supply chain with one manufacturer one retailer and onecarrier in the presence of complex uncertainties It is assumedthat demand 119863119905 follows a normal distribution ie 119863119905 sim119873(1205831 12059021) And the upstreammanufacturerrsquos yield is unstableThere are twomethods (strategies) for the whole supply chainto cope with uncertain risks information sharing (IS) or not(NS) Under strategy IS real-time information on marketdemand and the retailerrsquos inventory is shared among allmembersTherefore the valuable information can be used byeach member to adjust own decision to adapt to the externalenvironment However demand and inventory informationis not shared under strategy NS where it is difficult to makedynamic decisions for some members The detailed channelstructures under two strategies are showed in Figure 1 Thewhole event of our model is dynamic which includes twostages
Stage 1 The whole supply chain jointly decides whether toshare information among all members IS or NS
Stage 2 Under the given strategy the second stage is madeup of multiple periods During each period the sequence ofevents is as follows
(1) At the beginning of each period the manufacturerforecasts an order in advance and completes produc-tion
(2) The demand is realized(3) The retailer firstly meets the back orders and market
demand through available inventory in hand Thenthe order point is adjusted through sharing informa-tionunder strategy IS but it is constant under strategyNS Lastly whether to place an order upstream isdecided Unmet demand will be delayed to nextperiod if the inventory is enough
(4) The transportation capability is forecasted in advanceby the carrier with shared information under strategyIS but it is constant under strategy NS
(5) If the manufacturer accepts the retailerrsquos order theorder is transported to the retailer by the carrier(when yield is not enough insufficient orders aredelayed until the next period) otherwise go to (6)
(6) Inventories of the manufacturer and the retailer arechecked and the leftovers will be still sold in nextperiods
(7) All members compute the total cost to prepare for thenext period
The parameters and variables used throughout the paperare defined in Table 1
32 The Retailer Agent
321 Retailerrsquos Behavior under Strategy NS Under strategyNS four tasks are completed in turn according to the
4 Discrete Dynamics in Nature and Society
Table 1 The decision variables and parameters in the model
Decision variablesNotation Description Notation Description119896119905 Transportation capacity 119910119905 Manufacturerrsquos order forecast119904119905 Order point
ParametersNotation Description Notation Description119905 Period 120587119904119888 Supply chainrsquos total cost120587119877 Retailerrsquos total cost 120587119872 Manufacturerrsquos total costMH Retailerrsquos in-transit inventory 119877119864119905 Retailerrsquos remaining inventoryMR Order received by the retailer 119901119906119899119894119904ℎ119905119903119886119899119904 Total delayed penalty cost of the carrierML Retailerrsquos in-transit inventory 120582 Yield risk factorI Retailerrsquos initial inventory 119861119874119872 Manufacturerrsquos back orderIE Retailerrsquos ending inventory H Manufacturerrsquos unit inventory holding costIP Retailerrsquos current inventory B Manufacturerrsquos unit short cost119861119874119877 Retailerrsquos back order 119868minus119872119890119899119889 Manufacturerrsquos ending inventory119871119879119877 Order process time 119868minus119872119904119905119886119903119905 Manufacturerrsquos initial inventoryS Themaximum inventory level sale Delivered order quantity of the manufacturer120575 The safety factor on inventory 119901119906119899119894119904ℎ119872 Total penalty cost of the manufacturerQ Retailerrsquos order quantity 119891 The cost of maintaining the transportation capacityh Retailerrsquos unit inventory holding cost 119887 Retailerrsquos unit short cost120587119905119903119886119899119904 Carrierrsquos total cost 119863 The actual market demand119871119879119905119903119886119899119904 Transportation time 120572 120573 System parameters119888119905119903119886119899119904 Unit transportation cost 1205831 Mean of uncertain market demand119888119901 Unit penalty cost due to transportation delay 1205901 Standard deviation of uncertain market demand1205833 Mean of order process time 1205902 Standard deviation of uncertain yield1205903 Standard deviation of order process time 1205904 Standard deviation of transportation time1205834 Mean of transportation time 119874119872 Fixed order cost of the manufacturer119874119877 Fixed order cost of the retailer
time sequence during each period inventory check demandfulfillment inventory management and cost compute
(1) Inventory Check Before demand is realized in each periodthe order quantity from upstream is ensured by the retailer
119872119877119905 = 119872119867119905minus119871119879 +119872119867119905minus119871119879minus1 (1)where119872119877119905 is the retailerrsquos order received from upstream inperiod 119905119872119867119905minus119871119879 (119872119867119905minus119871119879minus1) is the retailerrsquos order quantityin period t-LT (t-LT -1)
Then the initial inventory and in-transit inventory arerespectively updated
119872119871 119905 = 119872119871 119905minus1 minus119872119877119905 (3)where 119868119905 is the retailerrsquos initial inventory at the beginning ofperiod 119905 119868119864119905minus1 is the ending inventory in the last period t-1119872119871 119905 is the total in-transit inventory in period 119905(2) Demand Fulfillment The former back orders and marketdemand are met through available inventory
119877119864119905 = 119868119905 minus 119861119874119877119905minus1 minus 119863119905 119868119905 gt 119861119874119877119905minus1 + 1198631199050 119868119905 le 119861119874119877119905minus1 + 119863119905 (4)
where 119877119864119905 is the remaining inventory in period 119905 119861119874119877119905minus1 isthe retailerrsquos total delayed order in the last period t-1
(3) Inventory Management It is assumed that famous (119904119905 119878)inventory policy is used Similar to Axsater [32]
where 119904119905 is the order point 1199040 is the initial value of 119904119905 andit is a constant under strategy IS 119904119905 (119905 = 1 2 ) = 1199040 119878is the maximum inventory level and the initial inventory inthe first period 1198681 = 119878 120575 is the safety factor on inventory119887 is the retailerrsquos unit delayed cost ℎ is the unit inventoryholding cost119874119877 is the retailerrsquos ordering cost 119871119879119877 and 119871119879119905119903119886119899119904are respectively the lead time of the order process time andtransportation time which are random variables followingnormal distribution 119871119879119877 sim 119873(1205833 12059023) 119871119879119905119903119886119899119904 sim 119873(1205834 12059024)and 119871119879 = 1205833 + 1205834
Discrete Dynamics in Nature and Society 5
Inventory check
NS
Demandfulfillment
Order based on
Compute cost
Inventory check
Demandfulfillment
Order based on (sS) rule
(sS) rule
Compute cost
Update the order point s
IS
The three tasks are completed by the
manufacturer instead of the retailer
Figure 2 The retailerrsquos behavior under two strategies
The current inventory level 119868119875119905 is119868119875119905 = 119877119864119905 +119872119871 119905 (8)
The order quantity in this period is
119876119905 = 119878 minus 119868119875119905 119868119875119905 lt 1198780 119868119875119905 ge 119878 (9)
where 120587119873119878(119868119878)119877119905 is the retailerrsquos total cost under strategy NS(IS) in period t the first term is the total inventory holdingcost the second term is the total delayed cost due to unmetdemand the third term is the total carrierrsquos punishment costthe fourth term is the total manufacturerrsquos punishment costand the last term is the fixed order cost
322 Retailerrsquos Behavior under Strategy IS Under strategy ISinventory check and management are accomplished by themanufacturer in lieu of the retailer The detail is presented inSection 332 Other behaviors are the same as those understrategy NS The retailerrsquos behavior under two cases is shownin Figure 2
33 The Manufacturer Agent
331 Manufacturerrsquos Behavior under Strategy NS Understrategy NS the work of forecast and production demandfulfillment inventorymanagement and cost computation areconducted in turn
(1) Forecast and Production Because of a long lead time formany products forecast and production must be finishedbefore the selling season in order to respond to consumersrapidly Hence the mode of make-to-stock is adopted by themanufacturer
In most cases the manufacturer cannot know the marketdemand information clearly under strategy NS After allthere is a retailer between the manufacturer and consumersmarket Further it is often hard and costly to obtain completeinformation on uncertain market for a manufacturer Thusproduction quantity is forecasted based on orders from thedownstream retailer [6 10]
Similar to Teunter et al [10] the commonmoving averagemethod is utilized to forecast the order quantity after Nperiods The forecast is based on historical order quantities119876119895 (119895 = 119905 minus 1 119905 minus 2 1) from the retailer 119910119905 is the forecastvalue in period 119905 119910119905 is a constant 119910119900 when 119905 lt 119873
Then the production is competed It is assumed that themanufacturer is subjected to yield risk due to the uncertainproduction process The actual yield is 120582119910119905 The commonproportion model is used here to describe this randomphenomenon 120582 a multiplication factor is set to be a randomvariable following normal distribution 120582 sim 119873(1 12059022) [33]
6 Discrete Dynamics in Nature and Society
(2) Demand Fulfillment First initial inventory is updated inaccord with yield and the ending inventory in last period
119868minus119872119904119905119886119903119905119905 is the manufacturerrsquos initial inventory inperiod 119905 119868minus119872119890119899119889119905minus1 is the ending inventory in the last periodt-1
Then the demand is met
119904119886119897119890119905 = min (119868minus119872119904119905119886119903119905119905 119861119874119872119905minus1 + 119876119905) (15)
119904119886119897119890119905 is the actual fulfillment quantity in period t 119861119874119872119905minus1is the manufacturerrsquos total short order in the last period t-1
(3) Inventory Management The ending inventory and backorder are checked
119868 119872119890119899119889119905=
0 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 119876119905119868 119872119904119905119886119903119905119905 minus 119861119874119872119905minus1 minus 119876119905 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(16)
119861119874119872119905=
119861119874119872119905minus1 + 119876119905 minus 119868 119872119904119905119886119903119905119905 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 1198761199050 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(17)
119868 119872119890119899119889119905 are regarded as remaining inventories to be soldin next periods and short orders 119861119874119872119905 are delayed to fulfillin next periods
(4) Cost The total cost of the manufacturer in each period is
where 120587119873119878(119868119878)119872119905 is the manufacturerrsquos total cost under strategyIS (NS) in period 119905119867 is unit inventory holding cost 119861 is themanufacturerrsquos unit short cost Hence the first term is thetotal inventory holding cost the second term is the total shortcost the last term is the fixed order cost
332Manufacturerrsquos Behavior under Strategy IS Under strat-egy IS two behaviors are different from those under strategyNS
Firstly the order forecast is dependent on shared marketdemand data rather than the historical order quantities after119873 periods Likewise 119910119905 is a constant 119910119900 as 119905 lt 119873
Market demand information can be shared by the retailerunder strategy IS when the manufacturerrsquos production canbe forecasted in light of direct market demand rather thana retailerrsquos orders As a result of the famous bullwhip effect
[11] market demand information is more accurate for amanufacturer compared with the information on a retailerrsquoorders
Secondly the retailerrsquos inventory is specially managed bythe manufacturer (119904119905 119878) policy is still adopted under strategyIS Due to the shared information of market demand andinventory on the one hand the retailerrsquos order process timeis removed ie 119871119879119877 = 0 Thus the initial value of the orderpoint 1199040 = 1205831 sdot1205833+120575sdot1205833 sdot1205901 On the other hand the order point119904119905 can be adjusted dynamically after N periods to decreaseoperations cost 119904119905 = 1199040 if 119905 lt 119873The decision rule is as belowwhich is dependent on historical experience [21]
Δ119904 = 119904119905 minus 119904119905minus1 (21)
Only if |Δ119904| ge 120572 sdot 119904119905minus1 (0 lt 120572 lt 1) 119904119905 replaces 119904119905minus1 120572 is aconstant coefficient
The manufacturerrsquos behavior under two cases is shown inFigure 3
34 The Carrier Agent The manufacturerrsquos products aretransported by the carrier The delivery lead time is 119871119879119905119903119886119899119904which is assumed to follow the normal distribution 119871119879119905119903119886119899119904 sim119873(1205833 12059023 ) The transportation capacity 119896119905 is reserved beforeeach delivery which is a constant 1198960 under strategy NSThe cost of maintaining the transportation capacity is 119891 =120573119896119905 (0 lt 120573 lt 1) 120573 is the maintaining cost of unit capacity Ifthe freight volume is less than 119896119905 the delivery time is 119871119879119905119903119886119899119904otherwise the delivery time is 119871119879119905119903119886119899119904 +1 [4] and the delayedpunishment cost is
However the capacity 119896119905 is a dynamic decision variableunder strategy IS 119896119905 can be determined dynamically in lightof some shared information after119873 periods [21]
119896119905
=
1198960 119905 lt 119873119896119905minus1 + Δ119904 119905 ge 119873 and
where 120587119873119878(119868119878)119905119903119886119899119904119905 is the carrierrsquos total cost under strategy NS (IS)in period 119905 the first term is delayed punishment cost the
Discrete Dynamics in Nature and Society 7
Forecast based onhistorical orders
NS IS
Downstream order fulfillment
Own inventory check
Compute cost
Forecast based onhistorical demand
Manage the inventory of retailer
Downstream order fulfillment
Own inventorycheck
Compute cost
Market demand is realized
Downstream order fulfillment
Update new order point s
No order required
Meet party demand
No
Yes
Yes
Place an order No
No
Yes
helliphellip
）Ｎ ge ＄Ｎ
）０Ｎge Ｍ
Ｎge
Figure 3 The manufacturerrsquos behavior under two strategies
second term is capacity maintaining cost the third term isthe delivery cost
The carrierrsquos behavior under two cases is presented inFigure 4
Finally the supply chainrsquos total cost is examined which isthe cost sum of three members
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
2 Discrete Dynamics in Nature and Society
not willing to provide information to others due to environ-mental uncertainties After all they are afraid that the simpleinformation sharing behavior may not cope with the complexuncertainties [5] Given the different attitudes toward theinformation sharing under an uncertain situation this paperexplores how multiple external uncertainties influence infor-mation sharing strategy for a supply chain such as uncertaindemand supply and lead time
The actual supply chain is proved to be a dynamicsystem [8] The complex dynamics are mainly reflected intwo aspects On the one hand the supply chain exists in anextremely uncertain environment where almost all externalelements vary all the time On the other hand the membersin the supply chain are autonomous individuals Sometimesthey adjust own decisions to adapt the dynamic exteriorcircumstance We refer to this supply chain as a dynamicsupply chain To the best of our knowledge the static optimaldecision is mainly concentrated on in conventional supplychain studies which are difficult to reflect the dynamic char-acteristic of the supply chain Hence unlike most researchesthe supply chain dynamics in a changing market that consistsof multiple periods however is the focus of this paper
An information sharing strategy is investigated in a three-level supply chain where one manufacturer one carrier andone retailer are faced with uncertain demand yield and leadtime There are two choices for the supply chain to manageexternal uncertainties information sharing (denoted by IS)or not (denoted by NS) Based on the method of multiagentmodelling we compare the total cost of the supply chainand each member in two cases to address the followingissues (1) Can the information sharing be beneficial tothe whole supply chain and each member in a dynamicenvironment simultaneously (2) What are the impacts ofsome uncertain risk factors on the information sharingdecisions We obtain meaningful management implicationsFor instance it is rewarding for the upstream manufacturerto use the retailerrsquos shared information under uncertain yieldor demand whereas information sharing strategy may beabandoned for the whole supply chain and other channelmembers and the retailer can obtain more benefits with theincrease of transportation time uncertainty
2 Literature Review
This paper is related to the information sharing in the supplychain and multiagent modeling
From the perspective of participants in the supply chainthe information sharing literature can be classified into twostreams The partnership between sellers and buyers is widelydiscussed in the first stream For example Cachon and Fisher[9] Lee et al [6] and Teunter et al [10] studied the valueof information sharing between upstream and downstreamfirms It is concluded that sharing information is beneficial toboth parties only under certain conditions Dejonckheere etal [11] Chatfield et al [12] Ma and Ma [13] and Zhao et al[14] analyzed the impacts of many factors on the well-knownbullwhip effect Aviv [15] Fildes and Kingsman [16] Traperoet al [17] and Sanders andWan [18] focused on how forecasterrors affect the information sharing whereas our paper is
different from these researches in that a carrier is regarded asa key supply chainmember here Particularly we consider theimpact of a carrier on the information sharing in a multilevelsupply chain The second stream is about the collaborationwith the third party logistics The typical studies are fewFor instance Wen [19] explored how to forecast shipmentof the carrier based on shared information However ourwork places emphasis on identifying complex uncertainfactors influencing the information sharing strategy Tyan etal [4] and Wen [3] qualitatively described the frameworkand competitive advantage of collaborative transportationmanagement (CTM) To be different quantitative study oninformation sharing behavior for a supply chainwith a carrieris conducted in our paper Chan and Zhang [20] and Li andChan [21] investigated the benefits of CTM in the mode ofmake-to-order (MTO) which is somewhat similar to oursIt is found that CTM lowers the total cost and risk for thewhole supply chain Yet the cost of each member is notdiscussed As a matter of fact it is necessary to guaranteeeach memberrsquos benefit in the case of information sharing Incontrast with [20 21] our differences are mainly displayedin four aspects (1) not only total cost of the whole supplychain but also that of each individual is examined as well (2)the manufacturer exogenous in their researches is served asan adaptive agent here which is more in line with practicalcases (3) the supply chain here is a make-to-stock (MTS)production system rather than the MTO system namely themanufacturerrsquos order is based on forecast (4) the effects ofmultiple uncertain risks are taken into account
Multiagent modeling is also correlated with our workThe traditional approaches about operations optimizationare widely adopted in the supply chain management whichattempts seeking the optimal decision Instead this staticoptimal behavior is usually not in line with practical casesAfter all the supply chain is a complex adaptive systemwhere each member has to be confronted with an uncertainsituation In addition practical members across the supplychain are bounded rational [22] who are hard to acquirecomplete information and find the best decision due tothe own ability In most cases adaptive learning throughpast experience is the common method to make decisionsMultiagent modeling (MAM) is a powerful and populartool to solve the complex dynamic problem owing to thedistinct strengths [23] Consequently MAM is introducedto depict the dynamic and autonomous features that weprimarily focus onThere have been representative literatureson MAM For example Swaminathan [23] Long [24] andYu and Wong [25] construct a framework to explore thesupply chain network dynamics Dogan and Guner [26] andHe et al [27] discuss pricing and ordering policies underdemand uncertainty In addition some other problems suchas inventory strategies [28] products management [29] andscheduling [30 31] are examined by many scholars as well
To sum up this paper contributes to the literature inseveral aspects First unlike many literatures the carrier isconsidered as a crucial member in a supply chain The issueof whether to share information with an intermediary carrierin a supply chain is investigated Each partyrsquos cost especially isstudied in detail Second we further explore the motivation
Discrete Dynamics in Nature and Society 3
ManufacturerRetailer
Information sharing platform
material flow
Carrier
material flow
information flow information flow
information flow
information flow information flow
ManufacturerRetailer
material flow
Carrier
material flow
information flow information flow
Order forecast Capacity forecastOrder pointadjustment
Strategy NS
Market demand and the inventory
Market
Market
material flow
material flow
Uncertain yield
Uncertain yield
Uncertain transportation Uncertain demand
Uncertain transportation Uncertain demand
Strategy IS
Figure 1 Two strategies IS and NS
to share information under external uncertainties To be spe-cific the impacts of uncertain demand yield and lead timeon information sharing are discussed Lastly the complexsupply chainrsquos dynamic and adaptive natures are captured inthis paper In particular each member is capable of alteringown decisions in a dynamic environment
3 The Model
31 The Overall Structure and Problem Description Considera supply chain with one manufacturer one retailer and onecarrier in the presence of complex uncertainties It is assumedthat demand 119863119905 follows a normal distribution ie 119863119905 sim119873(1205831 12059021) And the upstreammanufacturerrsquos yield is unstableThere are twomethods (strategies) for the whole supply chainto cope with uncertain risks information sharing (IS) or not(NS) Under strategy IS real-time information on marketdemand and the retailerrsquos inventory is shared among allmembersTherefore the valuable information can be used byeach member to adjust own decision to adapt to the externalenvironment However demand and inventory informationis not shared under strategy NS where it is difficult to makedynamic decisions for some members The detailed channelstructures under two strategies are showed in Figure 1 Thewhole event of our model is dynamic which includes twostages
Stage 1 The whole supply chain jointly decides whether toshare information among all members IS or NS
Stage 2 Under the given strategy the second stage is madeup of multiple periods During each period the sequence ofevents is as follows
(1) At the beginning of each period the manufacturerforecasts an order in advance and completes produc-tion
(2) The demand is realized(3) The retailer firstly meets the back orders and market
demand through available inventory in hand Thenthe order point is adjusted through sharing informa-tionunder strategy IS but it is constant under strategyNS Lastly whether to place an order upstream isdecided Unmet demand will be delayed to nextperiod if the inventory is enough
(4) The transportation capability is forecasted in advanceby the carrier with shared information under strategyIS but it is constant under strategy NS
(5) If the manufacturer accepts the retailerrsquos order theorder is transported to the retailer by the carrier(when yield is not enough insufficient orders aredelayed until the next period) otherwise go to (6)
(6) Inventories of the manufacturer and the retailer arechecked and the leftovers will be still sold in nextperiods
(7) All members compute the total cost to prepare for thenext period
The parameters and variables used throughout the paperare defined in Table 1
32 The Retailer Agent
321 Retailerrsquos Behavior under Strategy NS Under strategyNS four tasks are completed in turn according to the
4 Discrete Dynamics in Nature and Society
Table 1 The decision variables and parameters in the model
Decision variablesNotation Description Notation Description119896119905 Transportation capacity 119910119905 Manufacturerrsquos order forecast119904119905 Order point
ParametersNotation Description Notation Description119905 Period 120587119904119888 Supply chainrsquos total cost120587119877 Retailerrsquos total cost 120587119872 Manufacturerrsquos total costMH Retailerrsquos in-transit inventory 119877119864119905 Retailerrsquos remaining inventoryMR Order received by the retailer 119901119906119899119894119904ℎ119905119903119886119899119904 Total delayed penalty cost of the carrierML Retailerrsquos in-transit inventory 120582 Yield risk factorI Retailerrsquos initial inventory 119861119874119872 Manufacturerrsquos back orderIE Retailerrsquos ending inventory H Manufacturerrsquos unit inventory holding costIP Retailerrsquos current inventory B Manufacturerrsquos unit short cost119861119874119877 Retailerrsquos back order 119868minus119872119890119899119889 Manufacturerrsquos ending inventory119871119879119877 Order process time 119868minus119872119904119905119886119903119905 Manufacturerrsquos initial inventoryS Themaximum inventory level sale Delivered order quantity of the manufacturer120575 The safety factor on inventory 119901119906119899119894119904ℎ119872 Total penalty cost of the manufacturerQ Retailerrsquos order quantity 119891 The cost of maintaining the transportation capacityh Retailerrsquos unit inventory holding cost 119887 Retailerrsquos unit short cost120587119905119903119886119899119904 Carrierrsquos total cost 119863 The actual market demand119871119879119905119903119886119899119904 Transportation time 120572 120573 System parameters119888119905119903119886119899119904 Unit transportation cost 1205831 Mean of uncertain market demand119888119901 Unit penalty cost due to transportation delay 1205901 Standard deviation of uncertain market demand1205833 Mean of order process time 1205902 Standard deviation of uncertain yield1205903 Standard deviation of order process time 1205904 Standard deviation of transportation time1205834 Mean of transportation time 119874119872 Fixed order cost of the manufacturer119874119877 Fixed order cost of the retailer
time sequence during each period inventory check demandfulfillment inventory management and cost compute
(1) Inventory Check Before demand is realized in each periodthe order quantity from upstream is ensured by the retailer
119872119877119905 = 119872119867119905minus119871119879 +119872119867119905minus119871119879minus1 (1)where119872119877119905 is the retailerrsquos order received from upstream inperiod 119905119872119867119905minus119871119879 (119872119867119905minus119871119879minus1) is the retailerrsquos order quantityin period t-LT (t-LT -1)
Then the initial inventory and in-transit inventory arerespectively updated
119872119871 119905 = 119872119871 119905minus1 minus119872119877119905 (3)where 119868119905 is the retailerrsquos initial inventory at the beginning ofperiod 119905 119868119864119905minus1 is the ending inventory in the last period t-1119872119871 119905 is the total in-transit inventory in period 119905(2) Demand Fulfillment The former back orders and marketdemand are met through available inventory
119877119864119905 = 119868119905 minus 119861119874119877119905minus1 minus 119863119905 119868119905 gt 119861119874119877119905minus1 + 1198631199050 119868119905 le 119861119874119877119905minus1 + 119863119905 (4)
where 119877119864119905 is the remaining inventory in period 119905 119861119874119877119905minus1 isthe retailerrsquos total delayed order in the last period t-1
(3) Inventory Management It is assumed that famous (119904119905 119878)inventory policy is used Similar to Axsater [32]
where 119904119905 is the order point 1199040 is the initial value of 119904119905 andit is a constant under strategy IS 119904119905 (119905 = 1 2 ) = 1199040 119878is the maximum inventory level and the initial inventory inthe first period 1198681 = 119878 120575 is the safety factor on inventory119887 is the retailerrsquos unit delayed cost ℎ is the unit inventoryholding cost119874119877 is the retailerrsquos ordering cost 119871119879119877 and 119871119879119905119903119886119899119904are respectively the lead time of the order process time andtransportation time which are random variables followingnormal distribution 119871119879119877 sim 119873(1205833 12059023) 119871119879119905119903119886119899119904 sim 119873(1205834 12059024)and 119871119879 = 1205833 + 1205834
Discrete Dynamics in Nature and Society 5
Inventory check
NS
Demandfulfillment
Order based on
Compute cost
Inventory check
Demandfulfillment
Order based on (sS) rule
(sS) rule
Compute cost
Update the order point s
IS
The three tasks are completed by the
manufacturer instead of the retailer
Figure 2 The retailerrsquos behavior under two strategies
The current inventory level 119868119875119905 is119868119875119905 = 119877119864119905 +119872119871 119905 (8)
The order quantity in this period is
119876119905 = 119878 minus 119868119875119905 119868119875119905 lt 1198780 119868119875119905 ge 119878 (9)
where 120587119873119878(119868119878)119877119905 is the retailerrsquos total cost under strategy NS(IS) in period t the first term is the total inventory holdingcost the second term is the total delayed cost due to unmetdemand the third term is the total carrierrsquos punishment costthe fourth term is the total manufacturerrsquos punishment costand the last term is the fixed order cost
322 Retailerrsquos Behavior under Strategy IS Under strategy ISinventory check and management are accomplished by themanufacturer in lieu of the retailer The detail is presented inSection 332 Other behaviors are the same as those understrategy NS The retailerrsquos behavior under two cases is shownin Figure 2
33 The Manufacturer Agent
331 Manufacturerrsquos Behavior under Strategy NS Understrategy NS the work of forecast and production demandfulfillment inventorymanagement and cost computation areconducted in turn
(1) Forecast and Production Because of a long lead time formany products forecast and production must be finishedbefore the selling season in order to respond to consumersrapidly Hence the mode of make-to-stock is adopted by themanufacturer
In most cases the manufacturer cannot know the marketdemand information clearly under strategy NS After allthere is a retailer between the manufacturer and consumersmarket Further it is often hard and costly to obtain completeinformation on uncertain market for a manufacturer Thusproduction quantity is forecasted based on orders from thedownstream retailer [6 10]
Similar to Teunter et al [10] the commonmoving averagemethod is utilized to forecast the order quantity after Nperiods The forecast is based on historical order quantities119876119895 (119895 = 119905 minus 1 119905 minus 2 1) from the retailer 119910119905 is the forecastvalue in period 119905 119910119905 is a constant 119910119900 when 119905 lt 119873
Then the production is competed It is assumed that themanufacturer is subjected to yield risk due to the uncertainproduction process The actual yield is 120582119910119905 The commonproportion model is used here to describe this randomphenomenon 120582 a multiplication factor is set to be a randomvariable following normal distribution 120582 sim 119873(1 12059022) [33]
6 Discrete Dynamics in Nature and Society
(2) Demand Fulfillment First initial inventory is updated inaccord with yield and the ending inventory in last period
119868minus119872119904119905119886119903119905119905 is the manufacturerrsquos initial inventory inperiod 119905 119868minus119872119890119899119889119905minus1 is the ending inventory in the last periodt-1
Then the demand is met
119904119886119897119890119905 = min (119868minus119872119904119905119886119903119905119905 119861119874119872119905minus1 + 119876119905) (15)
119904119886119897119890119905 is the actual fulfillment quantity in period t 119861119874119872119905minus1is the manufacturerrsquos total short order in the last period t-1
(3) Inventory Management The ending inventory and backorder are checked
119868 119872119890119899119889119905=
0 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 119876119905119868 119872119904119905119886119903119905119905 minus 119861119874119872119905minus1 minus 119876119905 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(16)
119861119874119872119905=
119861119874119872119905minus1 + 119876119905 minus 119868 119872119904119905119886119903119905119905 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 1198761199050 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(17)
119868 119872119890119899119889119905 are regarded as remaining inventories to be soldin next periods and short orders 119861119874119872119905 are delayed to fulfillin next periods
(4) Cost The total cost of the manufacturer in each period is
where 120587119873119878(119868119878)119872119905 is the manufacturerrsquos total cost under strategyIS (NS) in period 119905119867 is unit inventory holding cost 119861 is themanufacturerrsquos unit short cost Hence the first term is thetotal inventory holding cost the second term is the total shortcost the last term is the fixed order cost
332Manufacturerrsquos Behavior under Strategy IS Under strat-egy IS two behaviors are different from those under strategyNS
Firstly the order forecast is dependent on shared marketdemand data rather than the historical order quantities after119873 periods Likewise 119910119905 is a constant 119910119900 as 119905 lt 119873
Market demand information can be shared by the retailerunder strategy IS when the manufacturerrsquos production canbe forecasted in light of direct market demand rather thana retailerrsquos orders As a result of the famous bullwhip effect
[11] market demand information is more accurate for amanufacturer compared with the information on a retailerrsquoorders
Secondly the retailerrsquos inventory is specially managed bythe manufacturer (119904119905 119878) policy is still adopted under strategyIS Due to the shared information of market demand andinventory on the one hand the retailerrsquos order process timeis removed ie 119871119879119877 = 0 Thus the initial value of the orderpoint 1199040 = 1205831 sdot1205833+120575sdot1205833 sdot1205901 On the other hand the order point119904119905 can be adjusted dynamically after N periods to decreaseoperations cost 119904119905 = 1199040 if 119905 lt 119873The decision rule is as belowwhich is dependent on historical experience [21]
Δ119904 = 119904119905 minus 119904119905minus1 (21)
Only if |Δ119904| ge 120572 sdot 119904119905minus1 (0 lt 120572 lt 1) 119904119905 replaces 119904119905minus1 120572 is aconstant coefficient
The manufacturerrsquos behavior under two cases is shown inFigure 3
34 The Carrier Agent The manufacturerrsquos products aretransported by the carrier The delivery lead time is 119871119879119905119903119886119899119904which is assumed to follow the normal distribution 119871119879119905119903119886119899119904 sim119873(1205833 12059023 ) The transportation capacity 119896119905 is reserved beforeeach delivery which is a constant 1198960 under strategy NSThe cost of maintaining the transportation capacity is 119891 =120573119896119905 (0 lt 120573 lt 1) 120573 is the maintaining cost of unit capacity Ifthe freight volume is less than 119896119905 the delivery time is 119871119879119905119903119886119899119904otherwise the delivery time is 119871119879119905119903119886119899119904 +1 [4] and the delayedpunishment cost is
However the capacity 119896119905 is a dynamic decision variableunder strategy IS 119896119905 can be determined dynamically in lightof some shared information after119873 periods [21]
119896119905
=
1198960 119905 lt 119873119896119905minus1 + Δ119904 119905 ge 119873 and
where 120587119873119878(119868119878)119905119903119886119899119904119905 is the carrierrsquos total cost under strategy NS (IS)in period 119905 the first term is delayed punishment cost the
Discrete Dynamics in Nature and Society 7
Forecast based onhistorical orders
NS IS
Downstream order fulfillment
Own inventory check
Compute cost
Forecast based onhistorical demand
Manage the inventory of retailer
Downstream order fulfillment
Own inventorycheck
Compute cost
Market demand is realized
Downstream order fulfillment
Update new order point s
No order required
Meet party demand
No
Yes
Yes
Place an order No
No
Yes
helliphellip
）Ｎ ge ＄Ｎ
）０Ｎge Ｍ
Ｎge
Figure 3 The manufacturerrsquos behavior under two strategies
second term is capacity maintaining cost the third term isthe delivery cost
The carrierrsquos behavior under two cases is presented inFigure 4
Finally the supply chainrsquos total cost is examined which isthe cost sum of three members
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
Discrete Dynamics in Nature and Society 3
ManufacturerRetailer
Information sharing platform
material flow
Carrier
material flow
information flow information flow
information flow
information flow information flow
ManufacturerRetailer
material flow
Carrier
material flow
information flow information flow
Order forecast Capacity forecastOrder pointadjustment
Strategy NS
Market demand and the inventory
Market
Market
material flow
material flow
Uncertain yield
Uncertain yield
Uncertain transportation Uncertain demand
Uncertain transportation Uncertain demand
Strategy IS
Figure 1 Two strategies IS and NS
to share information under external uncertainties To be spe-cific the impacts of uncertain demand yield and lead timeon information sharing are discussed Lastly the complexsupply chainrsquos dynamic and adaptive natures are captured inthis paper In particular each member is capable of alteringown decisions in a dynamic environment
3 The Model
31 The Overall Structure and Problem Description Considera supply chain with one manufacturer one retailer and onecarrier in the presence of complex uncertainties It is assumedthat demand 119863119905 follows a normal distribution ie 119863119905 sim119873(1205831 12059021) And the upstreammanufacturerrsquos yield is unstableThere are twomethods (strategies) for the whole supply chainto cope with uncertain risks information sharing (IS) or not(NS) Under strategy IS real-time information on marketdemand and the retailerrsquos inventory is shared among allmembersTherefore the valuable information can be used byeach member to adjust own decision to adapt to the externalenvironment However demand and inventory informationis not shared under strategy NS where it is difficult to makedynamic decisions for some members The detailed channelstructures under two strategies are showed in Figure 1 Thewhole event of our model is dynamic which includes twostages
Stage 1 The whole supply chain jointly decides whether toshare information among all members IS or NS
Stage 2 Under the given strategy the second stage is madeup of multiple periods During each period the sequence ofevents is as follows
(1) At the beginning of each period the manufacturerforecasts an order in advance and completes produc-tion
(2) The demand is realized(3) The retailer firstly meets the back orders and market
demand through available inventory in hand Thenthe order point is adjusted through sharing informa-tionunder strategy IS but it is constant under strategyNS Lastly whether to place an order upstream isdecided Unmet demand will be delayed to nextperiod if the inventory is enough
(4) The transportation capability is forecasted in advanceby the carrier with shared information under strategyIS but it is constant under strategy NS
(5) If the manufacturer accepts the retailerrsquos order theorder is transported to the retailer by the carrier(when yield is not enough insufficient orders aredelayed until the next period) otherwise go to (6)
(6) Inventories of the manufacturer and the retailer arechecked and the leftovers will be still sold in nextperiods
(7) All members compute the total cost to prepare for thenext period
The parameters and variables used throughout the paperare defined in Table 1
32 The Retailer Agent
321 Retailerrsquos Behavior under Strategy NS Under strategyNS four tasks are completed in turn according to the
4 Discrete Dynamics in Nature and Society
Table 1 The decision variables and parameters in the model
Decision variablesNotation Description Notation Description119896119905 Transportation capacity 119910119905 Manufacturerrsquos order forecast119904119905 Order point
ParametersNotation Description Notation Description119905 Period 120587119904119888 Supply chainrsquos total cost120587119877 Retailerrsquos total cost 120587119872 Manufacturerrsquos total costMH Retailerrsquos in-transit inventory 119877119864119905 Retailerrsquos remaining inventoryMR Order received by the retailer 119901119906119899119894119904ℎ119905119903119886119899119904 Total delayed penalty cost of the carrierML Retailerrsquos in-transit inventory 120582 Yield risk factorI Retailerrsquos initial inventory 119861119874119872 Manufacturerrsquos back orderIE Retailerrsquos ending inventory H Manufacturerrsquos unit inventory holding costIP Retailerrsquos current inventory B Manufacturerrsquos unit short cost119861119874119877 Retailerrsquos back order 119868minus119872119890119899119889 Manufacturerrsquos ending inventory119871119879119877 Order process time 119868minus119872119904119905119886119903119905 Manufacturerrsquos initial inventoryS Themaximum inventory level sale Delivered order quantity of the manufacturer120575 The safety factor on inventory 119901119906119899119894119904ℎ119872 Total penalty cost of the manufacturerQ Retailerrsquos order quantity 119891 The cost of maintaining the transportation capacityh Retailerrsquos unit inventory holding cost 119887 Retailerrsquos unit short cost120587119905119903119886119899119904 Carrierrsquos total cost 119863 The actual market demand119871119879119905119903119886119899119904 Transportation time 120572 120573 System parameters119888119905119903119886119899119904 Unit transportation cost 1205831 Mean of uncertain market demand119888119901 Unit penalty cost due to transportation delay 1205901 Standard deviation of uncertain market demand1205833 Mean of order process time 1205902 Standard deviation of uncertain yield1205903 Standard deviation of order process time 1205904 Standard deviation of transportation time1205834 Mean of transportation time 119874119872 Fixed order cost of the manufacturer119874119877 Fixed order cost of the retailer
time sequence during each period inventory check demandfulfillment inventory management and cost compute
(1) Inventory Check Before demand is realized in each periodthe order quantity from upstream is ensured by the retailer
119872119877119905 = 119872119867119905minus119871119879 +119872119867119905minus119871119879minus1 (1)where119872119877119905 is the retailerrsquos order received from upstream inperiod 119905119872119867119905minus119871119879 (119872119867119905minus119871119879minus1) is the retailerrsquos order quantityin period t-LT (t-LT -1)
Then the initial inventory and in-transit inventory arerespectively updated
119872119871 119905 = 119872119871 119905minus1 minus119872119877119905 (3)where 119868119905 is the retailerrsquos initial inventory at the beginning ofperiod 119905 119868119864119905minus1 is the ending inventory in the last period t-1119872119871 119905 is the total in-transit inventory in period 119905(2) Demand Fulfillment The former back orders and marketdemand are met through available inventory
119877119864119905 = 119868119905 minus 119861119874119877119905minus1 minus 119863119905 119868119905 gt 119861119874119877119905minus1 + 1198631199050 119868119905 le 119861119874119877119905minus1 + 119863119905 (4)
where 119877119864119905 is the remaining inventory in period 119905 119861119874119877119905minus1 isthe retailerrsquos total delayed order in the last period t-1
(3) Inventory Management It is assumed that famous (119904119905 119878)inventory policy is used Similar to Axsater [32]
where 119904119905 is the order point 1199040 is the initial value of 119904119905 andit is a constant under strategy IS 119904119905 (119905 = 1 2 ) = 1199040 119878is the maximum inventory level and the initial inventory inthe first period 1198681 = 119878 120575 is the safety factor on inventory119887 is the retailerrsquos unit delayed cost ℎ is the unit inventoryholding cost119874119877 is the retailerrsquos ordering cost 119871119879119877 and 119871119879119905119903119886119899119904are respectively the lead time of the order process time andtransportation time which are random variables followingnormal distribution 119871119879119877 sim 119873(1205833 12059023) 119871119879119905119903119886119899119904 sim 119873(1205834 12059024)and 119871119879 = 1205833 + 1205834
Discrete Dynamics in Nature and Society 5
Inventory check
NS
Demandfulfillment
Order based on
Compute cost
Inventory check
Demandfulfillment
Order based on (sS) rule
(sS) rule
Compute cost
Update the order point s
IS
The three tasks are completed by the
manufacturer instead of the retailer
Figure 2 The retailerrsquos behavior under two strategies
The current inventory level 119868119875119905 is119868119875119905 = 119877119864119905 +119872119871 119905 (8)
The order quantity in this period is
119876119905 = 119878 minus 119868119875119905 119868119875119905 lt 1198780 119868119875119905 ge 119878 (9)
where 120587119873119878(119868119878)119877119905 is the retailerrsquos total cost under strategy NS(IS) in period t the first term is the total inventory holdingcost the second term is the total delayed cost due to unmetdemand the third term is the total carrierrsquos punishment costthe fourth term is the total manufacturerrsquos punishment costand the last term is the fixed order cost
322 Retailerrsquos Behavior under Strategy IS Under strategy ISinventory check and management are accomplished by themanufacturer in lieu of the retailer The detail is presented inSection 332 Other behaviors are the same as those understrategy NS The retailerrsquos behavior under two cases is shownin Figure 2
33 The Manufacturer Agent
331 Manufacturerrsquos Behavior under Strategy NS Understrategy NS the work of forecast and production demandfulfillment inventorymanagement and cost computation areconducted in turn
(1) Forecast and Production Because of a long lead time formany products forecast and production must be finishedbefore the selling season in order to respond to consumersrapidly Hence the mode of make-to-stock is adopted by themanufacturer
In most cases the manufacturer cannot know the marketdemand information clearly under strategy NS After allthere is a retailer between the manufacturer and consumersmarket Further it is often hard and costly to obtain completeinformation on uncertain market for a manufacturer Thusproduction quantity is forecasted based on orders from thedownstream retailer [6 10]
Similar to Teunter et al [10] the commonmoving averagemethod is utilized to forecast the order quantity after Nperiods The forecast is based on historical order quantities119876119895 (119895 = 119905 minus 1 119905 minus 2 1) from the retailer 119910119905 is the forecastvalue in period 119905 119910119905 is a constant 119910119900 when 119905 lt 119873
Then the production is competed It is assumed that themanufacturer is subjected to yield risk due to the uncertainproduction process The actual yield is 120582119910119905 The commonproportion model is used here to describe this randomphenomenon 120582 a multiplication factor is set to be a randomvariable following normal distribution 120582 sim 119873(1 12059022) [33]
6 Discrete Dynamics in Nature and Society
(2) Demand Fulfillment First initial inventory is updated inaccord with yield and the ending inventory in last period
119868minus119872119904119905119886119903119905119905 is the manufacturerrsquos initial inventory inperiod 119905 119868minus119872119890119899119889119905minus1 is the ending inventory in the last periodt-1
Then the demand is met
119904119886119897119890119905 = min (119868minus119872119904119905119886119903119905119905 119861119874119872119905minus1 + 119876119905) (15)
119904119886119897119890119905 is the actual fulfillment quantity in period t 119861119874119872119905minus1is the manufacturerrsquos total short order in the last period t-1
(3) Inventory Management The ending inventory and backorder are checked
119868 119872119890119899119889119905=
0 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 119876119905119868 119872119904119905119886119903119905119905 minus 119861119874119872119905minus1 minus 119876119905 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(16)
119861119874119872119905=
119861119874119872119905minus1 + 119876119905 minus 119868 119872119904119905119886119903119905119905 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 1198761199050 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(17)
119868 119872119890119899119889119905 are regarded as remaining inventories to be soldin next periods and short orders 119861119874119872119905 are delayed to fulfillin next periods
(4) Cost The total cost of the manufacturer in each period is
where 120587119873119878(119868119878)119872119905 is the manufacturerrsquos total cost under strategyIS (NS) in period 119905119867 is unit inventory holding cost 119861 is themanufacturerrsquos unit short cost Hence the first term is thetotal inventory holding cost the second term is the total shortcost the last term is the fixed order cost
332Manufacturerrsquos Behavior under Strategy IS Under strat-egy IS two behaviors are different from those under strategyNS
Firstly the order forecast is dependent on shared marketdemand data rather than the historical order quantities after119873 periods Likewise 119910119905 is a constant 119910119900 as 119905 lt 119873
Market demand information can be shared by the retailerunder strategy IS when the manufacturerrsquos production canbe forecasted in light of direct market demand rather thana retailerrsquos orders As a result of the famous bullwhip effect
[11] market demand information is more accurate for amanufacturer compared with the information on a retailerrsquoorders
Secondly the retailerrsquos inventory is specially managed bythe manufacturer (119904119905 119878) policy is still adopted under strategyIS Due to the shared information of market demand andinventory on the one hand the retailerrsquos order process timeis removed ie 119871119879119877 = 0 Thus the initial value of the orderpoint 1199040 = 1205831 sdot1205833+120575sdot1205833 sdot1205901 On the other hand the order point119904119905 can be adjusted dynamically after N periods to decreaseoperations cost 119904119905 = 1199040 if 119905 lt 119873The decision rule is as belowwhich is dependent on historical experience [21]
Δ119904 = 119904119905 minus 119904119905minus1 (21)
Only if |Δ119904| ge 120572 sdot 119904119905minus1 (0 lt 120572 lt 1) 119904119905 replaces 119904119905minus1 120572 is aconstant coefficient
The manufacturerrsquos behavior under two cases is shown inFigure 3
34 The Carrier Agent The manufacturerrsquos products aretransported by the carrier The delivery lead time is 119871119879119905119903119886119899119904which is assumed to follow the normal distribution 119871119879119905119903119886119899119904 sim119873(1205833 12059023 ) The transportation capacity 119896119905 is reserved beforeeach delivery which is a constant 1198960 under strategy NSThe cost of maintaining the transportation capacity is 119891 =120573119896119905 (0 lt 120573 lt 1) 120573 is the maintaining cost of unit capacity Ifthe freight volume is less than 119896119905 the delivery time is 119871119879119905119903119886119899119904otherwise the delivery time is 119871119879119905119903119886119899119904 +1 [4] and the delayedpunishment cost is
However the capacity 119896119905 is a dynamic decision variableunder strategy IS 119896119905 can be determined dynamically in lightof some shared information after119873 periods [21]
119896119905
=
1198960 119905 lt 119873119896119905minus1 + Δ119904 119905 ge 119873 and
where 120587119873119878(119868119878)119905119903119886119899119904119905 is the carrierrsquos total cost under strategy NS (IS)in period 119905 the first term is delayed punishment cost the
Discrete Dynamics in Nature and Society 7
Forecast based onhistorical orders
NS IS
Downstream order fulfillment
Own inventory check
Compute cost
Forecast based onhistorical demand
Manage the inventory of retailer
Downstream order fulfillment
Own inventorycheck
Compute cost
Market demand is realized
Downstream order fulfillment
Update new order point s
No order required
Meet party demand
No
Yes
Yes
Place an order No
No
Yes
helliphellip
）Ｎ ge ＄Ｎ
）０Ｎge Ｍ
Ｎge
Figure 3 The manufacturerrsquos behavior under two strategies
second term is capacity maintaining cost the third term isthe delivery cost
The carrierrsquos behavior under two cases is presented inFigure 4
Finally the supply chainrsquos total cost is examined which isthe cost sum of three members
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
4 Discrete Dynamics in Nature and Society
Table 1 The decision variables and parameters in the model
Decision variablesNotation Description Notation Description119896119905 Transportation capacity 119910119905 Manufacturerrsquos order forecast119904119905 Order point
ParametersNotation Description Notation Description119905 Period 120587119904119888 Supply chainrsquos total cost120587119877 Retailerrsquos total cost 120587119872 Manufacturerrsquos total costMH Retailerrsquos in-transit inventory 119877119864119905 Retailerrsquos remaining inventoryMR Order received by the retailer 119901119906119899119894119904ℎ119905119903119886119899119904 Total delayed penalty cost of the carrierML Retailerrsquos in-transit inventory 120582 Yield risk factorI Retailerrsquos initial inventory 119861119874119872 Manufacturerrsquos back orderIE Retailerrsquos ending inventory H Manufacturerrsquos unit inventory holding costIP Retailerrsquos current inventory B Manufacturerrsquos unit short cost119861119874119877 Retailerrsquos back order 119868minus119872119890119899119889 Manufacturerrsquos ending inventory119871119879119877 Order process time 119868minus119872119904119905119886119903119905 Manufacturerrsquos initial inventoryS Themaximum inventory level sale Delivered order quantity of the manufacturer120575 The safety factor on inventory 119901119906119899119894119904ℎ119872 Total penalty cost of the manufacturerQ Retailerrsquos order quantity 119891 The cost of maintaining the transportation capacityh Retailerrsquos unit inventory holding cost 119887 Retailerrsquos unit short cost120587119905119903119886119899119904 Carrierrsquos total cost 119863 The actual market demand119871119879119905119903119886119899119904 Transportation time 120572 120573 System parameters119888119905119903119886119899119904 Unit transportation cost 1205831 Mean of uncertain market demand119888119901 Unit penalty cost due to transportation delay 1205901 Standard deviation of uncertain market demand1205833 Mean of order process time 1205902 Standard deviation of uncertain yield1205903 Standard deviation of order process time 1205904 Standard deviation of transportation time1205834 Mean of transportation time 119874119872 Fixed order cost of the manufacturer119874119877 Fixed order cost of the retailer
time sequence during each period inventory check demandfulfillment inventory management and cost compute
(1) Inventory Check Before demand is realized in each periodthe order quantity from upstream is ensured by the retailer
119872119877119905 = 119872119867119905minus119871119879 +119872119867119905minus119871119879minus1 (1)where119872119877119905 is the retailerrsquos order received from upstream inperiod 119905119872119867119905minus119871119879 (119872119867119905minus119871119879minus1) is the retailerrsquos order quantityin period t-LT (t-LT -1)
Then the initial inventory and in-transit inventory arerespectively updated
119872119871 119905 = 119872119871 119905minus1 minus119872119877119905 (3)where 119868119905 is the retailerrsquos initial inventory at the beginning ofperiod 119905 119868119864119905minus1 is the ending inventory in the last period t-1119872119871 119905 is the total in-transit inventory in period 119905(2) Demand Fulfillment The former back orders and marketdemand are met through available inventory
119877119864119905 = 119868119905 minus 119861119874119877119905minus1 minus 119863119905 119868119905 gt 119861119874119877119905minus1 + 1198631199050 119868119905 le 119861119874119877119905minus1 + 119863119905 (4)
where 119877119864119905 is the remaining inventory in period 119905 119861119874119877119905minus1 isthe retailerrsquos total delayed order in the last period t-1
(3) Inventory Management It is assumed that famous (119904119905 119878)inventory policy is used Similar to Axsater [32]
where 119904119905 is the order point 1199040 is the initial value of 119904119905 andit is a constant under strategy IS 119904119905 (119905 = 1 2 ) = 1199040 119878is the maximum inventory level and the initial inventory inthe first period 1198681 = 119878 120575 is the safety factor on inventory119887 is the retailerrsquos unit delayed cost ℎ is the unit inventoryholding cost119874119877 is the retailerrsquos ordering cost 119871119879119877 and 119871119879119905119903119886119899119904are respectively the lead time of the order process time andtransportation time which are random variables followingnormal distribution 119871119879119877 sim 119873(1205833 12059023) 119871119879119905119903119886119899119904 sim 119873(1205834 12059024)and 119871119879 = 1205833 + 1205834
Discrete Dynamics in Nature and Society 5
Inventory check
NS
Demandfulfillment
Order based on
Compute cost
Inventory check
Demandfulfillment
Order based on (sS) rule
(sS) rule
Compute cost
Update the order point s
IS
The three tasks are completed by the
manufacturer instead of the retailer
Figure 2 The retailerrsquos behavior under two strategies
The current inventory level 119868119875119905 is119868119875119905 = 119877119864119905 +119872119871 119905 (8)
The order quantity in this period is
119876119905 = 119878 minus 119868119875119905 119868119875119905 lt 1198780 119868119875119905 ge 119878 (9)
where 120587119873119878(119868119878)119877119905 is the retailerrsquos total cost under strategy NS(IS) in period t the first term is the total inventory holdingcost the second term is the total delayed cost due to unmetdemand the third term is the total carrierrsquos punishment costthe fourth term is the total manufacturerrsquos punishment costand the last term is the fixed order cost
322 Retailerrsquos Behavior under Strategy IS Under strategy ISinventory check and management are accomplished by themanufacturer in lieu of the retailer The detail is presented inSection 332 Other behaviors are the same as those understrategy NS The retailerrsquos behavior under two cases is shownin Figure 2
33 The Manufacturer Agent
331 Manufacturerrsquos Behavior under Strategy NS Understrategy NS the work of forecast and production demandfulfillment inventorymanagement and cost computation areconducted in turn
(1) Forecast and Production Because of a long lead time formany products forecast and production must be finishedbefore the selling season in order to respond to consumersrapidly Hence the mode of make-to-stock is adopted by themanufacturer
In most cases the manufacturer cannot know the marketdemand information clearly under strategy NS After allthere is a retailer between the manufacturer and consumersmarket Further it is often hard and costly to obtain completeinformation on uncertain market for a manufacturer Thusproduction quantity is forecasted based on orders from thedownstream retailer [6 10]
Similar to Teunter et al [10] the commonmoving averagemethod is utilized to forecast the order quantity after Nperiods The forecast is based on historical order quantities119876119895 (119895 = 119905 minus 1 119905 minus 2 1) from the retailer 119910119905 is the forecastvalue in period 119905 119910119905 is a constant 119910119900 when 119905 lt 119873
Then the production is competed It is assumed that themanufacturer is subjected to yield risk due to the uncertainproduction process The actual yield is 120582119910119905 The commonproportion model is used here to describe this randomphenomenon 120582 a multiplication factor is set to be a randomvariable following normal distribution 120582 sim 119873(1 12059022) [33]
6 Discrete Dynamics in Nature and Society
(2) Demand Fulfillment First initial inventory is updated inaccord with yield and the ending inventory in last period
119868minus119872119904119905119886119903119905119905 is the manufacturerrsquos initial inventory inperiod 119905 119868minus119872119890119899119889119905minus1 is the ending inventory in the last periodt-1
Then the demand is met
119904119886119897119890119905 = min (119868minus119872119904119905119886119903119905119905 119861119874119872119905minus1 + 119876119905) (15)
119904119886119897119890119905 is the actual fulfillment quantity in period t 119861119874119872119905minus1is the manufacturerrsquos total short order in the last period t-1
(3) Inventory Management The ending inventory and backorder are checked
119868 119872119890119899119889119905=
0 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 119876119905119868 119872119904119905119886119903119905119905 minus 119861119874119872119905minus1 minus 119876119905 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(16)
119861119874119872119905=
119861119874119872119905minus1 + 119876119905 minus 119868 119872119904119905119886119903119905119905 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 1198761199050 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(17)
119868 119872119890119899119889119905 are regarded as remaining inventories to be soldin next periods and short orders 119861119874119872119905 are delayed to fulfillin next periods
(4) Cost The total cost of the manufacturer in each period is
where 120587119873119878(119868119878)119872119905 is the manufacturerrsquos total cost under strategyIS (NS) in period 119905119867 is unit inventory holding cost 119861 is themanufacturerrsquos unit short cost Hence the first term is thetotal inventory holding cost the second term is the total shortcost the last term is the fixed order cost
332Manufacturerrsquos Behavior under Strategy IS Under strat-egy IS two behaviors are different from those under strategyNS
Firstly the order forecast is dependent on shared marketdemand data rather than the historical order quantities after119873 periods Likewise 119910119905 is a constant 119910119900 as 119905 lt 119873
Market demand information can be shared by the retailerunder strategy IS when the manufacturerrsquos production canbe forecasted in light of direct market demand rather thana retailerrsquos orders As a result of the famous bullwhip effect
[11] market demand information is more accurate for amanufacturer compared with the information on a retailerrsquoorders
Secondly the retailerrsquos inventory is specially managed bythe manufacturer (119904119905 119878) policy is still adopted under strategyIS Due to the shared information of market demand andinventory on the one hand the retailerrsquos order process timeis removed ie 119871119879119877 = 0 Thus the initial value of the orderpoint 1199040 = 1205831 sdot1205833+120575sdot1205833 sdot1205901 On the other hand the order point119904119905 can be adjusted dynamically after N periods to decreaseoperations cost 119904119905 = 1199040 if 119905 lt 119873The decision rule is as belowwhich is dependent on historical experience [21]
Δ119904 = 119904119905 minus 119904119905minus1 (21)
Only if |Δ119904| ge 120572 sdot 119904119905minus1 (0 lt 120572 lt 1) 119904119905 replaces 119904119905minus1 120572 is aconstant coefficient
The manufacturerrsquos behavior under two cases is shown inFigure 3
34 The Carrier Agent The manufacturerrsquos products aretransported by the carrier The delivery lead time is 119871119879119905119903119886119899119904which is assumed to follow the normal distribution 119871119879119905119903119886119899119904 sim119873(1205833 12059023 ) The transportation capacity 119896119905 is reserved beforeeach delivery which is a constant 1198960 under strategy NSThe cost of maintaining the transportation capacity is 119891 =120573119896119905 (0 lt 120573 lt 1) 120573 is the maintaining cost of unit capacity Ifthe freight volume is less than 119896119905 the delivery time is 119871119879119905119903119886119899119904otherwise the delivery time is 119871119879119905119903119886119899119904 +1 [4] and the delayedpunishment cost is
However the capacity 119896119905 is a dynamic decision variableunder strategy IS 119896119905 can be determined dynamically in lightof some shared information after119873 periods [21]
119896119905
=
1198960 119905 lt 119873119896119905minus1 + Δ119904 119905 ge 119873 and
where 120587119873119878(119868119878)119905119903119886119899119904119905 is the carrierrsquos total cost under strategy NS (IS)in period 119905 the first term is delayed punishment cost the
Discrete Dynamics in Nature and Society 7
Forecast based onhistorical orders
NS IS
Downstream order fulfillment
Own inventory check
Compute cost
Forecast based onhistorical demand
Manage the inventory of retailer
Downstream order fulfillment
Own inventorycheck
Compute cost
Market demand is realized
Downstream order fulfillment
Update new order point s
No order required
Meet party demand
No
Yes
Yes
Place an order No
No
Yes
helliphellip
）Ｎ ge ＄Ｎ
）０Ｎge Ｍ
Ｎge
Figure 3 The manufacturerrsquos behavior under two strategies
second term is capacity maintaining cost the third term isthe delivery cost
The carrierrsquos behavior under two cases is presented inFigure 4
Finally the supply chainrsquos total cost is examined which isthe cost sum of three members
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
where 120587119873119878(119868119878)119877119905 is the retailerrsquos total cost under strategy NS(IS) in period t the first term is the total inventory holdingcost the second term is the total delayed cost due to unmetdemand the third term is the total carrierrsquos punishment costthe fourth term is the total manufacturerrsquos punishment costand the last term is the fixed order cost
322 Retailerrsquos Behavior under Strategy IS Under strategy ISinventory check and management are accomplished by themanufacturer in lieu of the retailer The detail is presented inSection 332 Other behaviors are the same as those understrategy NS The retailerrsquos behavior under two cases is shownin Figure 2
33 The Manufacturer Agent
331 Manufacturerrsquos Behavior under Strategy NS Understrategy NS the work of forecast and production demandfulfillment inventorymanagement and cost computation areconducted in turn
(1) Forecast and Production Because of a long lead time formany products forecast and production must be finishedbefore the selling season in order to respond to consumersrapidly Hence the mode of make-to-stock is adopted by themanufacturer
In most cases the manufacturer cannot know the marketdemand information clearly under strategy NS After allthere is a retailer between the manufacturer and consumersmarket Further it is often hard and costly to obtain completeinformation on uncertain market for a manufacturer Thusproduction quantity is forecasted based on orders from thedownstream retailer [6 10]
Similar to Teunter et al [10] the commonmoving averagemethod is utilized to forecast the order quantity after Nperiods The forecast is based on historical order quantities119876119895 (119895 = 119905 minus 1 119905 minus 2 1) from the retailer 119910119905 is the forecastvalue in period 119905 119910119905 is a constant 119910119900 when 119905 lt 119873
Then the production is competed It is assumed that themanufacturer is subjected to yield risk due to the uncertainproduction process The actual yield is 120582119910119905 The commonproportion model is used here to describe this randomphenomenon 120582 a multiplication factor is set to be a randomvariable following normal distribution 120582 sim 119873(1 12059022) [33]
6 Discrete Dynamics in Nature and Society
(2) Demand Fulfillment First initial inventory is updated inaccord with yield and the ending inventory in last period
119868minus119872119904119905119886119903119905119905 is the manufacturerrsquos initial inventory inperiod 119905 119868minus119872119890119899119889119905minus1 is the ending inventory in the last periodt-1
Then the demand is met
119904119886119897119890119905 = min (119868minus119872119904119905119886119903119905119905 119861119874119872119905minus1 + 119876119905) (15)
119904119886119897119890119905 is the actual fulfillment quantity in period t 119861119874119872119905minus1is the manufacturerrsquos total short order in the last period t-1
(3) Inventory Management The ending inventory and backorder are checked
119868 119872119890119899119889119905=
0 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 119876119905119868 119872119904119905119886119903119905119905 minus 119861119874119872119905minus1 minus 119876119905 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(16)
119861119874119872119905=
119861119874119872119905minus1 + 119876119905 minus 119868 119872119904119905119886119903119905119905 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 1198761199050 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(17)
119868 119872119890119899119889119905 are regarded as remaining inventories to be soldin next periods and short orders 119861119874119872119905 are delayed to fulfillin next periods
(4) Cost The total cost of the manufacturer in each period is
where 120587119873119878(119868119878)119872119905 is the manufacturerrsquos total cost under strategyIS (NS) in period 119905119867 is unit inventory holding cost 119861 is themanufacturerrsquos unit short cost Hence the first term is thetotal inventory holding cost the second term is the total shortcost the last term is the fixed order cost
332Manufacturerrsquos Behavior under Strategy IS Under strat-egy IS two behaviors are different from those under strategyNS
Firstly the order forecast is dependent on shared marketdemand data rather than the historical order quantities after119873 periods Likewise 119910119905 is a constant 119910119900 as 119905 lt 119873
Market demand information can be shared by the retailerunder strategy IS when the manufacturerrsquos production canbe forecasted in light of direct market demand rather thana retailerrsquos orders As a result of the famous bullwhip effect
[11] market demand information is more accurate for amanufacturer compared with the information on a retailerrsquoorders
Secondly the retailerrsquos inventory is specially managed bythe manufacturer (119904119905 119878) policy is still adopted under strategyIS Due to the shared information of market demand andinventory on the one hand the retailerrsquos order process timeis removed ie 119871119879119877 = 0 Thus the initial value of the orderpoint 1199040 = 1205831 sdot1205833+120575sdot1205833 sdot1205901 On the other hand the order point119904119905 can be adjusted dynamically after N periods to decreaseoperations cost 119904119905 = 1199040 if 119905 lt 119873The decision rule is as belowwhich is dependent on historical experience [21]
Δ119904 = 119904119905 minus 119904119905minus1 (21)
Only if |Δ119904| ge 120572 sdot 119904119905minus1 (0 lt 120572 lt 1) 119904119905 replaces 119904119905minus1 120572 is aconstant coefficient
The manufacturerrsquos behavior under two cases is shown inFigure 3
34 The Carrier Agent The manufacturerrsquos products aretransported by the carrier The delivery lead time is 119871119879119905119903119886119899119904which is assumed to follow the normal distribution 119871119879119905119903119886119899119904 sim119873(1205833 12059023 ) The transportation capacity 119896119905 is reserved beforeeach delivery which is a constant 1198960 under strategy NSThe cost of maintaining the transportation capacity is 119891 =120573119896119905 (0 lt 120573 lt 1) 120573 is the maintaining cost of unit capacity Ifthe freight volume is less than 119896119905 the delivery time is 119871119879119905119903119886119899119904otherwise the delivery time is 119871119879119905119903119886119899119904 +1 [4] and the delayedpunishment cost is
However the capacity 119896119905 is a dynamic decision variableunder strategy IS 119896119905 can be determined dynamically in lightof some shared information after119873 periods [21]
119896119905
=
1198960 119905 lt 119873119896119905minus1 + Δ119904 119905 ge 119873 and
where 120587119873119878(119868119878)119905119903119886119899119904119905 is the carrierrsquos total cost under strategy NS (IS)in period 119905 the first term is delayed punishment cost the
Discrete Dynamics in Nature and Society 7
Forecast based onhistorical orders
NS IS
Downstream order fulfillment
Own inventory check
Compute cost
Forecast based onhistorical demand
Manage the inventory of retailer
Downstream order fulfillment
Own inventorycheck
Compute cost
Market demand is realized
Downstream order fulfillment
Update new order point s
No order required
Meet party demand
No
Yes
Yes
Place an order No
No
Yes
helliphellip
）Ｎ ge ＄Ｎ
）０Ｎge Ｍ
Ｎge
Figure 3 The manufacturerrsquos behavior under two strategies
second term is capacity maintaining cost the third term isthe delivery cost
The carrierrsquos behavior under two cases is presented inFigure 4
Finally the supply chainrsquos total cost is examined which isthe cost sum of three members
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
119868minus119872119904119905119886119903119905119905 is the manufacturerrsquos initial inventory inperiod 119905 119868minus119872119890119899119889119905minus1 is the ending inventory in the last periodt-1
Then the demand is met
119904119886119897119890119905 = min (119868minus119872119904119905119886119903119905119905 119861119874119872119905minus1 + 119876119905) (15)
119904119886119897119890119905 is the actual fulfillment quantity in period t 119861119874119872119905minus1is the manufacturerrsquos total short order in the last period t-1
(3) Inventory Management The ending inventory and backorder are checked
119868 119872119890119899119889119905=
0 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 119876119905119868 119872119904119905119886119903119905119905 minus 119861119874119872119905minus1 minus 119876119905 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(16)
119861119874119872119905=
119861119874119872119905minus1 + 119876119905 minus 119868 119872119904119905119886119903119905119905 119868 119872119904119905119886119903119905119905 le 119861119874119872119905minus1 + 1198761199050 119868 119872119904119905119886119903119905119905 gt 119861119874119872119905minus1 + 119876119905
(17)
119868 119872119890119899119889119905 are regarded as remaining inventories to be soldin next periods and short orders 119861119874119872119905 are delayed to fulfillin next periods
(4) Cost The total cost of the manufacturer in each period is
where 120587119873119878(119868119878)119872119905 is the manufacturerrsquos total cost under strategyIS (NS) in period 119905119867 is unit inventory holding cost 119861 is themanufacturerrsquos unit short cost Hence the first term is thetotal inventory holding cost the second term is the total shortcost the last term is the fixed order cost
332Manufacturerrsquos Behavior under Strategy IS Under strat-egy IS two behaviors are different from those under strategyNS
Firstly the order forecast is dependent on shared marketdemand data rather than the historical order quantities after119873 periods Likewise 119910119905 is a constant 119910119900 as 119905 lt 119873
Market demand information can be shared by the retailerunder strategy IS when the manufacturerrsquos production canbe forecasted in light of direct market demand rather thana retailerrsquos orders As a result of the famous bullwhip effect
[11] market demand information is more accurate for amanufacturer compared with the information on a retailerrsquoorders
Secondly the retailerrsquos inventory is specially managed bythe manufacturer (119904119905 119878) policy is still adopted under strategyIS Due to the shared information of market demand andinventory on the one hand the retailerrsquos order process timeis removed ie 119871119879119877 = 0 Thus the initial value of the orderpoint 1199040 = 1205831 sdot1205833+120575sdot1205833 sdot1205901 On the other hand the order point119904119905 can be adjusted dynamically after N periods to decreaseoperations cost 119904119905 = 1199040 if 119905 lt 119873The decision rule is as belowwhich is dependent on historical experience [21]
Δ119904 = 119904119905 minus 119904119905minus1 (21)
Only if |Δ119904| ge 120572 sdot 119904119905minus1 (0 lt 120572 lt 1) 119904119905 replaces 119904119905minus1 120572 is aconstant coefficient
The manufacturerrsquos behavior under two cases is shown inFigure 3
34 The Carrier Agent The manufacturerrsquos products aretransported by the carrier The delivery lead time is 119871119879119905119903119886119899119904which is assumed to follow the normal distribution 119871119879119905119903119886119899119904 sim119873(1205833 12059023 ) The transportation capacity 119896119905 is reserved beforeeach delivery which is a constant 1198960 under strategy NSThe cost of maintaining the transportation capacity is 119891 =120573119896119905 (0 lt 120573 lt 1) 120573 is the maintaining cost of unit capacity Ifthe freight volume is less than 119896119905 the delivery time is 119871119879119905119903119886119899119904otherwise the delivery time is 119871119879119905119903119886119899119904 +1 [4] and the delayedpunishment cost is
However the capacity 119896119905 is a dynamic decision variableunder strategy IS 119896119905 can be determined dynamically in lightof some shared information after119873 periods [21]
119896119905
=
1198960 119905 lt 119873119896119905minus1 + Δ119904 119905 ge 119873 and
where 120587119873119878(119868119878)119905119903119886119899119904119905 is the carrierrsquos total cost under strategy NS (IS)in period 119905 the first term is delayed punishment cost the
Discrete Dynamics in Nature and Society 7
Forecast based onhistorical orders
NS IS
Downstream order fulfillment
Own inventory check
Compute cost
Forecast based onhistorical demand
Manage the inventory of retailer
Downstream order fulfillment
Own inventorycheck
Compute cost
Market demand is realized
Downstream order fulfillment
Update new order point s
No order required
Meet party demand
No
Yes
Yes
Place an order No
No
Yes
helliphellip
）Ｎ ge ＄Ｎ
）０Ｎge Ｍ
Ｎge
Figure 3 The manufacturerrsquos behavior under two strategies
second term is capacity maintaining cost the third term isthe delivery cost
The carrierrsquos behavior under two cases is presented inFigure 4
Finally the supply chainrsquos total cost is examined which isthe cost sum of three members
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
where120587119873119878(119868119878)119904119888119905 is the supply chainrsquos total cost under strategyNS(IS) in period t
35 Algorithm
Step 1 119905 larr997888 1Step 2 Decision variables 1199040 1199100 1198960 and all exogenousparameters are initialized
Step 3 The manufacturer determines an order 119910119905 based onforecast
Step 4 Market demand 119863119905 is randomly realized
Step 5 The retailer firstly fulfills the former back orders andmarket demandThen the order point 119904119905 is updated accordingto formulas (20) and (21) under strategy IS however 119904119905 = 1199040
under strategy NS Lastly the retailer computes the orderquantity 119876119905Step 6 The transportation capability 119896119905 is adjusted accordingto formula (23) under strategy IS otherwise 119896119905 = 1198960 understrategy NS
Step 7 The products are transported to the retailer by thecarrier
Step 8 The total costs 120587119873119878(119868119878)119904119888119905 120587119873119878(119868119878)119905119903119886119899119904119905 120587119873119878(119868119878)119872119905 120587119873119878(119868119878)119877119905 arecomputed
Step 9 Enter next period (119905 larr997888 119905 + 1) and go to Step 3 untiltermination
Step 10 Compare the average cost of each member and thewhole supply chain under cases IS and NS
4 Simulation Experiments and Analysis
In this section the simulation experiments are firstlydesigned Then the effects of uncertain risks on the costs ofsupply chain members and information sharing strategy arestudied
Parameters of the experiments are set as Table 2 Sim-ulation experiments are conducted on the Eclipse platform
8 Discrete Dynamics in Nature and Society
Examine freight
NS
Compute cost
IS
Lead time is
Delayed penalty cost
Yes
NoExamine freight
Compute cost
Lead time is
Delayed penalty cost
Yes
No
YesNo
Adujst the
Examine current period t
volume salet
volume salet
Lead time is LTtrans
saletlekt
saletlekt
LTtrans+1
LTtrans+1
Lead time is LTtrans
tgeN
capacity kt
Figure 4 The carrierrsquos behavior under two strategies
Table 2 Thevalues of important parameters in experiments
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
with Java codes Experiments are carried out considering allparameters withmultiple values This combination method isused in the literature [34 35]The results in following figuresare shown on average Each simulation is run 100 times withdifferent random seeds and each time lasts for 500 periods togive each agent abundant time to learn historical experiences
010 015 020 025 030 035
500
600
700
800
900
Total cost of the manufacturer
IS
NS
The vertical gap the value of information sharing (IS)
2
Figure 5 Yield uncertainty versus the manufacturerrsquos costs undertwo cases
41 The Impacts of Uncertain Risks on the Channel Members
Observation 1 Under uncertain yield or demand strategy ISis a preferable choice for the manufacturer however it is notalways beneficial for other members to adopt IS
Firstly the effects of uncertain yield and demand onthe manufacturerrsquos costs under two strategies are explainedin Figures 5 and 6 respectively Strategy IS contributes tothe reduction of manufacturerrsquos cost under yield or demanduncertainty and the value of IS enlarges while the yield
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
Discrete Dynamics in Nature and Society 9
10 20 30 40 50 600
500
1000
1500
Total cost of the manufacturer
NS
IS
1
Figure 6 Demand uncertainty versus the manufacturerrsquos costs under two cases
010 015 020 025 030 035300
400
500
600
700
800
900
1000
Total cost of the retailer
NS
IS
A 2
Figure 7 Yield uncertainty versus the retailerrsquos costs under two cases
(demand) uncertainty increases The manufacturerrsquos forecastin each period is derived from the retailerrsquos past orders understrategy NS As a result of the bullwhip effect a crucialfactor for cost the manufacturerrsquos forecast is larger thanactual demand of the retailer However the retailerrsquos stockis managed by the manufacturer under strategy IS wherethe order process time is deleted and manufacturerrsquos forecastis based on market demand rather than retailerrsquos ordersTherefore the bullwhip effect is mitigated and inventoryholding cost and short cost are cut down Naturally it isbeneficial for the manufacturer to use the retailerrsquos sharedinformation However it is not the case for the retailer andthe carrier
Then the impacts of uncertain yield and demand on theretailerrsquos costs are studied Observed from Figures 7 and 8strategy IS is profitable for the retailer only when the yieldor demand uncertainty is not large But the cost gap is small
when yield or demand uncertainty is large Taking advantageof sharing information inventory forecast accuracy can beguaranteed if yield or demand uncertainty is not great Thusthe retailerrsquos inventory holding cost and delayed short costdecrease Yet forecast result is affected seriously if uncertaintyvalue is more than a threshold (1205901 gt 119860 119900119903 1205902 gt 119860)It is difficult to control these unnecessary costs incurredby risks Thus unlike the manufacturer strategy IS is notalways superior to the other for the retailer The value ofIS is not obvious as demand or yield uncertainty is largenamely information sharing should not be applied under thecircumstance
The impacts of yield demand and transportation timeuncertainties on the carrierrsquos costs are studied as well Similarto Figures 7 and 8 forecast accuracy is considered as asignificant element to trade off whether to share informationHence sometimes strategy IS is not better than NS for the
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
10 Discrete Dynamics in Nature and Society
10 20 30 40 50 60
300
600
900
1200
Total cost of the retailer
NS
IS
A 1
Figure 8 Demand uncertainty versus the retailerrsquos costs under two cases
Total cost of the retailer
NS
IS
0 1 2 3 4 5 6 70
1000
2000
3000
4
Figure 9 Transportation uncertainty versus the retailerrsquos costs under two cases
carrier If the uncertainties are large information sharingis not sensible Because of the similarity these details areomitted
Observation 2 A higher transportation time uncertaintyreduces the total cost of the retailer
Figure 9 illustrates how the uncertainty of transporta-tion time affects the retailerrsquos costs Counterintuitively theretailerrsquos total cost lowers with the transportation time uncer-tainty The uncertain transportation time is regarded as asignificant cause for the retailerrsquos stockout crisis Marketdemand fill rate decreases because of the increasing uncer-tainty which further gives rise to the more delayed short costfor the retailer However the penalty cost of the carrier dueto delayed delivery is enhanced as well while transportation
time becomes more uncertain Hence the retailerrsquos total costfinally decreases instead in that the carrierrsquos penalty cost theretailer obtains offsets increasing short cost
42 The Impacts of Uncertain Risks on the Supply Chain
Observation 3 Information sharing is not always beneficialto the whole supply chain under uncertain yield (demand)Strategy IS should be given up when yield (demand) uncer-tainty is large
The impact of yield uncertainty on the supply chaincosts under two cases are presented in Figure 10 Whenyield uncertainty is not large the value of strategy IS isevident otherwise strategy IS is worse than NS Channelmembers use shared information to adjust decisions and
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
Discrete Dynamics in Nature and Society 11
005 010 015 020 025 030 035 040
2100
2800
3500
Total cost of the supply chain
NS
IS
2
Figure 10 Yield uncertainty versus the supply chainrsquos costs under two cases
The total cost of supply chain
NS
IS
0 1 2 3 4 5 6 71000
1500
2000
3
Figure 11 Order process uncertainty versus the supply chainrsquos costs under two cases
adapt to environment dynamically under strategy IS whichsaves unnecessary costs caused by unstable yield if theseuncertainties are not large However it is not easy to controlthe risk when uncertainty is large in that forecast accuracyand quality is cut down Naturally the value of informationsharing is gradually weakening with the increase of yielduncertainty The result is similar to that of the demanduncertainty Therefore strategy IS should only be adopted bythe supply chain when external yield (demand) uncertaintyis not large Otherwise information sharing behavior shouldbe avoided
Observation 4 The cost caused by order process uncertaintycan be mitigated obviously under strategy IS but the advan-tage of strategy IS is not evident in terms of transportationtime uncertainty
The relationship between ordering process uncertaintyand supply chain costs is showed in Figure 11 The costunder strategy IS is smaller than that under NS Orderingprocess is a redundant activity under strategy NS whichincreases the total lead time and the retailerrsquos inventoryrisk Nevertheless the retailerrsquos inventory is managed by theupstream manufacturer under strategy IS Ordering processis omitted so total lead time and short cost decrease Hencethe negative impact of ordering process uncertainty canbe reduced if strategy IS is utilized especially under highuncertainty level It is profitable for the whole supply chainto share information when the ordering process time exists
The effect of transportation time uncertainty on supplychain costs is depicted in Figure 12 First it is clear thatunstable transportation time increases the supply chainrsquos
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
12 Discrete Dynamics in Nature and Society
Total cost of the supply chain
NS
IS
0 1 2 3 4 5 6
1000
1500
2000
2500
4
Figure 12 Transportation uncertainty versus the supply chainrsquoscosts under two cases
operations cost owing to the internal risk Moreover whilethe cost is less for strategy IS the value of IS is not remarkableAfter all the uncertainty in transport cannot be eliminatedin the spite of shared information Consequently it is hard tocontrol the risk caused by uncertain transportation
5 Conclusions
This paper studies an information sharing strategy in amultilevel supply chain with one manufacturer one carrierand one retailer where all members have to be confrontedwith uncertain yield demand and lead time in a complexmultiperiod environment Two strategies can be adoptedto react to multiple uncertainties IS or NS Each memberis regarded as an adaptive agent where decisions can beadjusted in each period to dynamically adapt to the externalsituation The costs of supply chain and channel membersunder two strategies are contrasted and the effects of yielddemand and lead time uncertainties on the two strategiesare investigated We find (i) strategy IS is optimal for theupstreammanufacturer under uncertain yield or demand (ii)but for the whole supply chain the retailer and the carrierstrategy IS is not always the suitable choice information shar-ing should be avoided when demand yield or transportationtime uncertainty is large (iii) the increase of transportationtime uncertainty benefits the retailer (iv) for the wholesupply chain the cost from ordering process uncertainty iscut down evidently through sharing information however itis not easy to mitigate the uncertain transportation risk withsharing information
There are several directions for future research First themanufacturerrsquos capacity is infinite This assumption could berelaxed to study a more complex case where the manufac-turer may be faced with capacity crisis Second it is worthstudying the impact of other decision adjustment methods oninformation sharing behavior Third market and inventoryinformation are shared among the supply chain members inthis paper but the yield risk upstream is not sharedThe factorcan be further considered and studied
Data Availability
My data is public
Conflicts of Interest
The authors declare that they have no conflicts of interest
References
[1] L Li ldquoInformation sharing in a supply chain with horizontalcompetitionrdquoManagement Science vol 48 no 9 pp 1196ndash12122002
[2] M A Darwish and O M Odah ldquoVendor managed inventorymodel for single-vendormulti-retailer supply chainsrdquoEuropeanJournal of Operational Research vol 204 no 3 pp 473ndash4842010
[3] Y-H Wen ldquoImpact of collaborative transportation manage-ment on logistics capability and competitive advantage for thecarrierrdquo Transportation Journal vol 51 no 4 pp 452ndash473 2012
[4] J C Tyan F KWang and T Du ldquoApplying collaborative trans-portation management models in global third-party logisticsrdquoInternational Journal of Computer Integrated Manufacturingvol 16 no 4-5 pp 283ndash291 2003
[5] Q Qi and Q Zhang ldquoResearch on information sharing risk insupply chain managementrdquo in Proceedings of the 4th Interna-tional Conference on Wireless Communications Networking andMobile Computing WiCOM rsquo08 pp 1ndash6 IEEE 2008
[6] H L Lee K C So andC S Tang ldquoValue of information sharingin a two-level supply chainrdquoManagement Science vol 46 no 5pp 626ndash643 2000
[7] Z Yu H Yan and T C E Cheng ldquoBenefits of informationsharingwith supply chain partnershipsrdquo IndustrialManagementand Data Systems vol 101 no 3 pp 114ndash121 2001
[8] A Surana S Kumara M Greaves and U N RaghavanldquoSupply-chain networks a complex adaptive systems perspec-tiverdquo International Journal of Production Research vol 43 no20 pp 4235ndash4265 2005
[9] G P Cachon and M Fisher ldquoSupply chain inventory man-agement and the value of shared informationrdquo ManagementScience vol 46 no 8 pp 1032ndash1048 2000
[10] R H Teunter M Z Babai J A Bokhorst and A A SyntetosldquoRevisiting the value of information sharing in two-stage supplychainsrdquo European Journal of Operational Research vol 270 no3 pp 1044ndash1052 2018
[11] J Dejonckheere S M Disney M R Lambrecht and D RTowill ldquoMeasuring and avoiding the bullwhip effect a controltheoretic approachrdquo European Journal of Operational Researchvol 147 no 3 pp 567ndash590 2003
[12] D C Chatfield J G Kim T P Harrison and J C Hayya ldquoThebullwhip effectmdashimpact of stochastic lead time informationquality and information sharing a simulation studyrdquo Produc-tion Engineering Research and Development vol 13 no 4 pp340ndash353 2004
[13] J Ma and X Ma ldquoMeasure of the bullwhip effect consideringthe market competition between two retailersrdquo InternationalJournal of Production Research vol 55 no 2 pp 313ndash326 2017
[14] Y Zhao Y Cao H Li et al ldquoBullwhip effect mitigation of greensupply chain optimization in electronics industryrdquo Journal ofCleaner Production vol 180 pp 888ndash912 2018
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
Numerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisNumerical AnalysisAdvances inAdvances in Discrete Dynamics in
Nature and SocietyHindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom
Dierential EquationsInternational Journal of
Volume 2018
Hindawiwwwhindawicom Volume 2018
Decision SciencesAdvances in
Hindawiwwwhindawicom Volume 2018
AnalysisInternational Journal of
Hindawiwwwhindawicom Volume 2018
Stochastic AnalysisInternational Journal of
Submit your manuscripts atwwwhindawicom
Discrete Dynamics in Nature and Society 13
[15] Y Aviv ldquoOn the benefits of collaborative forecasting part-nerships between retailers and manufacturersrdquo ManagementScience vol 53 no 5 pp 777ndash794 2007
[16] R Fildes and B Kingsman ldquoIncorporating demand uncertaintyand forecast error in supply chain planning modelsrdquo Journalof the Operational Research Society vol 62 no 3 pp 483ndash5002011
[17] J R Trapero N Kourentzes and R Fildes ldquoImpact of infor-mation exchange on supplier forecasting performancerdquo Omega vol 40 no 6 pp 738ndash747 2012
[18] N Sanders and X Wan ldquoMitigating forecast errors fromproduct variety through information sharingrdquo InternationalJournal of Production Research vol 56 no 12 pp 1ndash12 2018
[19] Y-HWen ldquoShipment forecasting for supply chain collaborativetransportation management using grey models with grey num-bersrdquoTransportation Planning and Technology vol 34 no 6 pp605ndash624 2011
[20] F T S Chan and T Zhang ldquoThe impact of collaborativetransportation management on supply chain performance asimulation approachrdquo Expert Systems with Applications vol 38no 3 pp 2319ndash2329 2011
[21] J Li and F T S Chan ldquoThe impact of collaborative transporta-tion management on demand disruption of manufacturingsupply chainsrdquo International Journal of Production Research vol50 no 19 pp 5635ndash5650 2012
[22] H A Simon ldquoTheories of bounded rationalityrdquo Decision andOrganization vol 1 no 1 pp 161ndash176 1972
[23] J M Swaminathan S F Smith and N M Sadeh ldquoModelingsupply chain dynamics a multiagent approachrdquo Decision Sci-ences vol 29 no 3 pp 607ndash631 1998
[24] Q Long ldquoThree-dimensional-flow model of agent-based com-putational experiment for complex supply network evolutionrdquoExpert Systems with Applications vol 42 no 5 pp 2525ndash25372015
[25] C Yu and T N Wong ldquoA multi-agent architecture for multi-product supplier selection in consideration of the synergybetween productsrdquo International Journal of Production Re-search vol 53 no 20 pp 6059ndash6082 2015
[26] I Dogan and A R Guner ldquoA reinforcement learning approachto competitive ordering and pricing problemrdquo Expert Systemswith Applications vol 32 no 1 pp 39ndash48 2015
[27] Z He SWang and T C E Cheng ldquoCompetition and evolutioninmulti-product supply chains An agent-based retailer modelrdquoInternational Journal of Production Economics vol 146 no 1 pp325ndash336 2013
[28] B Ponte E Sierra D de la Fuente and J Lozano ldquoExploringthe interaction of inventory policies across the supply chain anagent-based approachrdquo Computers amp Operations Research vol78 pp 335ndash348 2017
[29] I Giannoccaro and A Nair ldquoExamining the roles of productcomplexity andmanager behavior on product design decisionsan agent-based study using NK simulationrdquo IEEE Transactionson Engineering Management vol 63 no 2 pp 237ndash247 2016
[30] S Liu W H Wu C C Kang et al ldquoA single-machine two-agent scheduling problem by a branch-and-bound and threesimulated annealing algorithmsrdquo Discrete Dynamics in Natureand Society vol 2015 Article ID 681854 8 pages 2015
[31] L Wan ldquoTwo-agent scheduling tominimize the maximum costwith position-dependent jobsrdquoDiscreteDynamics inNature andSociety vol 2015 Article ID 932680 4 pages 2015
[32] S Axsater ldquoUsing the deterministic EOQ formula in stochasticinventory controlrdquoManagement Science vol 42 no 6 pp 830ndash834 1996
[33] F Lu H Xu P Chen and S X Zhu ldquoJoint pricing and pro-duction decisions with yield uncertainty and downconversionrdquoInternational Journal of Production Economics vol 197 pp 52ndash62 2018
[34] Z Liu ldquoEquilibrium analysis of capacity allocation withdemand competitionrdquo Naval Research Logistics (NRL) vol 59no 3-4 pp 254ndash265 2012
[35] K Cattani W Gilland H S Heese and J SwaminathanldquoBoiling frogs pricing strategies for a manufacturer addinga direct channel that competes with the traditional channelrdquoProduction Engineering Research and Development vol 15 no1 pp 40ndash56 2006
Hindawiwwwhindawicom Volume 2018
MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Mathematical Problems in Engineering
Applied MathematicsJournal of
Hindawiwwwhindawicom Volume 2018
Probability and StatisticsHindawiwwwhindawicom Volume 2018
Journal of
Hindawiwwwhindawicom Volume 2018
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawiwwwhindawicom Volume 2018
OptimizationJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Engineering Mathematics
International Journal of
Hindawiwwwhindawicom Volume 2018
Operations ResearchAdvances in
Journal of
Hindawiwwwhindawicom Volume 2018
Function SpacesAbstract and Applied AnalysisHindawiwwwhindawicom Volume 2018
International Journal of Mathematics and Mathematical Sciences
![Dynamic Spectrum Sharing by Opportunistic Spectrum Access ...epubs.surrey.ac.uk/807078/1/[150113] FinalThesis(HLee).pdf · Dynamic Spectrum Sharing by Opportunistic Spectrum Access](https://static.documents.pub/doc/80x56/5f2219dc6e2e315ceb610936/dynamic-spectrum-sharing-by-opportunistic-spectrum-access-epubs-150113-finalthesishleepdf.jpg)
