Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics Koji Kitazume, Candidate for Master of Environmental Management and Master of Business Administration Dr. Jay S. Golden, Advisor
December 2012
Master’s project submitted in partial fulfillment of the requirements for the Master of Environmental Management degree in the Nicholas School of the Environment of Duke University
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
i
Abstract
Nowadays, many logistics managers confront tradeoffs among keeping costs low, delivering
goods on time and reducing carbon footprint. In shipping finished goods from a manufacturing
plant in Asia to a distribution center in the eastern United States, how should a logistics manager
define and choose his preferred route and modes of transportation, taking into account the
potentially conflicting priorities?
This study explored a case of REI, an outdoor apparel brand/retailer, facing such a decision-
making question regarding its inbound logistics from the Port of Shanghai to its distribution center
in Bedford, Pennsylvania and approached it as a multiple objective problem. 15 possible intermodal
freight transportation routes with different attributes in terms of shipping costs, transit time and
greenhouse gas emissions were identified and associated data were collected. The preferred route
was derived by employing a simple additive model of preferences, using a pricing out method to
assess tradeoff weights and computing the overall utility of each alternative.
This framework quantified and visualized how the logistics manager’s choice is affected by
his preferences and the tradeoffs he is willing to make, thereby demonstrating its potential as a
practical aid for decision-making at the intersection of business and the environment. Accuracy of
the model used in this study could be improved by addressing uncertain data and omitted scope.
Furthermore, a versatile platform loaded and maintained with accurate and consistent data on
shipping costs, transit time and GHG emissions, covering multipoint-to-multipoint intermodal
freight transportation routes, could benefit shippers widely by enabling informed decision-making
to enhance their business and environmental performance.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
ii
Executive Summary
This study applied a multiple objective analysis approach for decision-making on the
selection of an intermodal freight transportation route for REI’s inbound logistics. 15 possible
intermodal freight transportation routes were identified to ship one full 40-foot standard container
load of a pant product from the Port of Shanghai to the distribution center in Bedford, Pennsylvania.
On all routes, the freight is transported from the Port of Shanghai to a US marine port via ocean. The
routes can be broadly-divided into four groups based on the location of their landing port. In
particular, Routes 1 through 3 can be grouped as the Pacific Northwest group, Routes 4 through 10
the California group, Routes 11 through 13 the South Atlantic group, and Routes 14 and 15 the Mid
Atlantic group. All groups but the Mid Atlantic use rail from the landing port to one of the
intermodal rail terminals within 360 miles from the Bedford distribution center. The remaining
segments of the routes use truck. With such variation in transportation modes and distance, the 15
alternatives present different attributes in terms of shipping costs, transit time and GHG emissions.
Table ES-1 summarizes the attributes for each route.
Route Costs
(US$)
Transit Time
(d:hh:mm)
GHG Emissions (kg CO2)
1 7,902 23:04:08 2,110 2 7,949 23:10:43 2,154 3 8,029 22:23:33 2,199 4 8,149 24:20:56 2,250 5 8,196 25:03:31 2,294 6 8,276 24:16:21 2,338 7 8,547 19:01:28 2,281 8 7,782 19:18:17 2,086 9 7,862 18:22:26 2,165
10 7,903 13:13:06 2,171 11 7,337 28:09:04 2,521 12 6,616 28:20:15 2,430 13 6,492 29:14:22 2,429 14 5,999 32:04:14 2,412 15 6,573 41:14:26 2,339
Table ES-1: Routes and Modes of Transportation for Analysis
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
iii
The objectives of this decision-making question on inbound transportation route selection
are to keep the three attributes of costs, transit time and emissions all low. Overall utility of each
route was computed using the following simple additive model of preferences:
For this purpose, outcomes of attributes described in Table ES-1 were converted to utilities
on a scale of 0 to 1 proportionately, with the best outcome of the attribute under consideration
being a 1 and the worst being a 0. Table ES-2 lists the converted utilities.
Route Costs
Transit Time GHG Emissions
1 0.25 0.66 0.95 2 0.23 0.65 0.84 3 0.20 0.66 0.74 4 0.16 0.60 0.62 5 0.14 0.59 0.52 6 0.11 0.60 0.42 7 0.00 0.80 0.55 8 0.30 0.78 1.00 9 0.27 0.81 0.82
10 0.25 1.00 0.81 11 0.47 0.47 0.00 12 0.76 0.45 0.21 13 0.81 0.43 0.21 14 1.00 0.34 0.25 15 0.77 0.00 0.42
Table ES-2: Utility Scores of Attributes
Assessment of tradeoff weights used the pricing out method. Table ES-3 presents the weight
assignments derived with the given assumptions.
Costs Transit Time GHG Emissions
Weight Table ES-3: Weights of Attributes
Overall utility of each route was computed by substituting the converted utilities and
derived weight assignments into the simple additive model of preferences equation. Table ES-4
presents the results.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
iv
Route Costs Transit Time GHG Emissions
Overall Utility Overall Rank
1 0.14 0.15 0.20 0.49 8 2 0.13 0.14 0.18 0.45 9 3 0.12 0.15 0.15 0.42 10 4 0.09 0.13 0.13 0.35 12 5 0.08 0.13 0.11 0.32 13 6 0.06 0.13 0.09 0.28 15 7 0.00 0.18 0.12 0.29 14 8 0.17 0.17 0.21 0.55 4 9 0.15 0.18 0.17 0.50 7
10 0.14 0.22 0.17 0.53 5 11 0.27 0.11 0.00 0.37 11 12 0.43 0.10 0.04 0.58 3 13 0.46 0.10 0.04 0.60 2 14 0.57 0.07 0.05 0.70 1 15 0.44 0.00 0.09 0.53 6
Table ES-4: Weighted and Overall Utility Scores and Ranking
Route 14 scored the highest overall utility. Therefore, this is logically the best route to ship
one full 40 foot standard container load of the pant product from the Port of Shanghai to the
Bedford DC, based on the given assumptions.
Sensitivity analysis of the tradeoff weight assessment was performed by increasing the
weight of one attribute while holding all other variables constant. Figure ES-1 illustrates how the
overall utility scores vary as the weight on the emissions attribute is increased. The preferred
alternative shifts from Route 14 to Route 8.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
v
Figure ES-1: Sensitivity Analysis by Increasing Weight of Emissions Attribute
The same was done on the transit time attribute. Figure ES-2 illustrates the results. The
preferred alternative shifts from Route 14 to Route 10 as the weight on transit time is increased.
Figure ES-2: Sensitivity Analysis by Increasing Weight of Time Attribute
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Emissions
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Time
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
vi
Table of Contents
Abstract ......................................................................................................................................................................................... i Executive Summary ................................................................................................................................................................ ii List of Tables ............................................................................................................................................................................. vi List of Figures .......................................................................................................................................................................... vii Glossary of Key Terms and Acronyms ......................................................................................................................... viii 1. Introduction ..................................................................................................................................................................... 1
1.1. Objective .................................................................................................................................................................. 1 2. Materials and Methods ................................................................................................................................................ 2
2.1. Preconditions ......................................................................................................................................................... 2 2.2. Stage One ................................................................................................................................................................. 3 2.3. Stage Two ................................................................................................................................................................ 4 2.4. Stage Three ............................................................................................................................................................. 5
3. Findings and Results ..................................................................................................................................................... 6 3.1. Freight Transportation in the Context of Global Warming ................................................................. 6 3.2. Stage One ................................................................................................................................................................. 8
3.2.1. Intermodal Transportation .................................................................................................................... 8 3.2.2. Ports ................................................................................................................................................................. 9 3.2.3. Ocean Freight ............................................................................................................................................. 12 3.2.4. Rail Freight .................................................................................................................................................. 14 3.2.5. Road Freight ............................................................................................................................................... 20 3.2.6. Routes and Modes of Transportation for Data Collection ....................................................... 21
3.3. Stage Two .............................................................................................................................................................. 23 3.3.1. Volume and Mass ...................................................................................................................................... 24 3.3.2. Distance ........................................................................................................................................................ 25 3.3.3. Emission Factors ....................................................................................................................................... 25 3.3.4. Costs ............................................................................................................................................................... 26 3.3.5. Transit Time ............................................................................................................................................... 28 3.3.6. Greenhouse Gas Emissions ................................................................................................................... 29 3.3.7. Routes and Modes of Transportation for Analysis ..................................................................... 29
3.4. Stage Three ........................................................................................................................................................... 30 3.4.1. Multiple Objective Decision-Making ................................................................................................. 30 3.4.2. Sensitivity Analysis .................................................................................................................................. 35 3.4.3. Other Types of Containers .................................................................................................................... 38
4. Discussion of Findings and Future Research ................................................................................................... 38 5. Conclusion ....................................................................................................................................................................... 41 References ................................................................................................................................................................................ 43 Attachments ............................................................................................................................................................................. 47
List of Tables
Table ES-1: Routes and Modes of Transportation for Analysis ............................................................................ ii Table ES-2: Utility Scores of Attributes ......................................................................................................................... iii Table ES-3: Weights of Attributes .................................................................................................................................... iii Table ES-4: Weighted and Overall Utility Scores and Ranking ............................................................................ iv Table 3-1: North America’s Top 25 Container Ports............................................................................................... 11 Table 3-2: World’s Top 20 Ocean Container Carriers ............................................................................................ 12
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
vii
Table 3-3: Maersk Transpacific Services ..................................................................................................................... 13 Table 3-4: Terminal Locations and Distance to Bedford Distribution Center .............................................. 20 Table 3-5: Routes and Modes of Transportation for Data Collection .............................................................. 23 Table 3-6: Distance per Route by Mode of Transportation .................................................................................. 25 Table 3-7: Emission Factors by Mode of Transportation...................................................................................... 26 Table 3-8: Total Costs per Route by Mode of Transportation ............................................................................. 27 Table 3-9: Transit Times per Route by Mode of Transportation ....................................................................... 28 Table 3-10: Total Greenhouse Gas Emissions per Route by Mode of Transportation .............................. 29 Table 3-11: Routes and Modes of Transportation for Analysis .......................................................................... 30 Table 3-12: Ranking of Attributes .................................................................................................................................. 31 Table 3-13: Range of Attributes....................................................................................................................................... 32 Table 3-14: Utility Scores of Attributes ........................................................................................................................ 33 Table 3-15: Weighted and Overall Utility Scores and Ranking........................................................................... 35 Table 3-16: Comparison of Results by Container Type ......................................................................................... 38
List of Figures
Figure ES-1: Sensitivity Analysis by Increasing Weight of Emissions Attribute ............................................ v Figure ES-2: Sensitivity Analysis by Increasing Weight of Time Attribute ...................................................... v Figure 3-1: GHG Emission Factor by Mode of Freight Transportation.............................................................. 7 Figure 3-2: US Top 25 Container Ports ......................................................................................................................... 10 Figure 3-3: BNSF System Map .......................................................................................................................................... 15 Figure 3-4: UP Systems Map ............................................................................................................................................. 16 Figure 3-5: CSXT System Map........................................................................................................................................... 17 Figure 3-6: NS System Map ................................................................................................................................................ 18 Figure 3-7: CN System Map ............................................................................................................................................... 19 Figure 3-8: CPR System Map ............................................................................................................................................. 20 Figure 3-9: Map of Terminal Locations and Bedford Distribution Center ..................................................... 21 Figure 3-10: Ranking of Attributes................................................................................................................................. 32 Figure 3-11: Sensitivity Analysis by Increasing Weight of Emissions Attribute ......................................... 36 Figure 3-12: Sensitivity Analysis by Increasing Weight of Time Attribute ................................................... 36 Figure 3-13: Sensitivity Analysis by Decreasing Cost per Mile of Rail Transportation ............................ 37
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
viii
Glossary of Key Terms and Acronyms
BNSF: BNSF Railway.
BSR: Business for Social Responsibility.
Carrier: A firm that transports goods or people via land, sea or air (Thomas Publishing Company, 2012).
Class I Railroad: A Class I railroad in the United States, or a Class I railway (also Class I rail carrier) in Canada, is one of the largest freight railroads, as classified based on operating revenue. The exact revenues required to be in each class have varied through the years, and they are now continuously adjusted for inflation. The threshold for a Class I Railroad in 2006 was $346.8 million (Canadian National Railway Company, 2012).
Class I Railway: See Class I Railroad.
CN: Canadian National Railway.
CO2: Carbon Dioxide.
CO2e: Carbon Dioxide-Equivalent.
CPR: Canadian Pacific Railway.
CSXT: CSX Transportation.
DC: Distribution Center – The warehouse facility which holds inventory from manufacturing pending distribution to the appropriate stores (Thomas Publishing Company, 2012).
Drayage: The service offered by a motor carrier for pick-up and delivery of ocean containers or rail containers (Thomas Publishing Company, 2012).
EPA: Environmental Protection Agency.
FEC: Florida East Coast Railway.
GHG: Greenhouse Gas.
GHG Protocol: The Greenhouse Gas Protocol Initiative.
ICTF: Intermodal Container Transfer Facility.
Inbound Logistics: The management of materials from suppliers and vendors into production processes or storage facilities (Thomas Publishing Company, 2012).
Intermodal Transportation: Transporting freight by using two or more transportation modes, such as by truck and rail or truck and oceangoing vessel (Thomas Publishing Company, 2012).
kg: Kilogram.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
ix
Maersk: Maersk Line – the global containerized division of the A.P. Moller – Maersk Group (Maersk Line).
mi: Mile.
Near-Dock: The ship-to-rail intermodal container transfer configuration that extends from the marine terminal and customs area to a nearby outside facility requiring drayage (Ashar & Swigart, 2007).
NGO: Non-Governmental Organization.
NS: Norfolk Southern Railway.
On-Dock: The ship-to-rail intermodal container transfer configuration that concludes within the marine terminal and customs area requiring no or minimal drayage (Ashar & Swigart, 2007).
Outbound Logistics: The process related to the movement and storage of products from the end of the production line to the end user (Thomas Publishing Company, 2012).
O-D Pair: Origin-Destination Pair.
Pallet: The platform which cartons are stacked on and then used for shipment or movement as a group. Pallets may be made of wood or composite materials (Thomas Publishing Company, 2012).
REI: Recreational Equipment, Inc.
Shipper: The party that tenders goods for transportation (Thomas Publishing Company, 2012).
Terminal: The end of a railroad or other transport route, or a station at such a point (Oxford University Press, 2012).
TEU: Twenty-Foot Equivalent Unit.
Transit Time: The total time that elapses between a shipment’s pickup and delivery (Thomas Publishing Company, 2012).
UP: Union Pacific Railroad.
US: United States.
WRI: World Resources Institute.
WTP: Willingness to Pay.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
1
1. Introduction
You are the inbound logistics manager of an American apparel brand. Specifically, you are in
charge of transporting finished goods from factories in Asia to your distribution centers in the
United States. You have been striving to achieve cost optimization targets while also meeting lead-
time expectations imposed by your merchandizing colleagues. On top of that, your company
recently decided to introduce greenhouse gas emissions reduction goals, and you will soon be
tasked to manage the carbon footprint of your inbound logistics as well.
You have two distribution centers in the US; one in the western region that covers the retail
stores in that half of the country, and another in the eastern region that covers those in the other
half. You are particularly interested in the inbound transportation to the eastern distribution center,
where you have the option to land your shipment from Asia on either the Pacific coast or the
Atlantic coast. How would you go about choosing the optimal inbound logistics, taking into account
the potentially conflicting priorities?
1.1. Objective
This study aims to provide a practical framework for addressing such a multiple objective
question of selecting the best intermodal freight transportation route for inbound logistics,
considering shipping costs, transit time and GHG emissions. The study is motivated by two real-
world challenges: First, while many freight carriers and NGOs offer a “carbon calculator” for moving
goods through transportation networks, there is not yet a de facto tool that allows end-to-end
inventorying of GHG emissions for routes across multiple carriers or different modes of
transportation. Second, as illustrated above, logistics managers are virtually never rewarded for
simply reducing their carbon footprint; rather, they are typically confronted with tradeoffs among
keeping costs low, delivering goods on time and cutting GHG emissions, which altogether could
make the decision-making of choosing the optimal inbound transportation route a challenge.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
2
As such, this study first identifies the financial, operational and environmental impacts of
inbound logistics for a specific origin-destination pair (i.e., transporting finished goods from the
factory in Asia to the distribution center in the US) via various routes and modes of transportation,
with a particular focus on intermodal freight transportation. Next, it attempts to demonstrate a
framework for decision-making involving such multiple objectives with tradeoffs. The hope is that
this study serves as a reference material for inbound logistics managers in their daily operations.
2. Materials and Methods
This study consists of three stages. The first stage will focus on identifying routes that are
available and suitable for analysis, through gaining a broad understanding of international
intermodal freight transportation. The second stage is spent on collecting cost, transit time and
GHG emissions data of the routes identified for detailed analysis. The third stage will be devoted to
modeling and analysis.
2.1. Preconditions
The apparel brand illustrated in the opening example is based on Recreational Equipment,
Inc., who agreed to support this study by sharing internal information and data. The outdoor
apparel brand/retailer has approximately $1.7 billion annual sales, 120 retail stores across the US,
and two distribution centers, one in the state of Washington covering the western region, and
another in Bedford, Pennsylvania covering the eastern region. This study will concentrate on REI’s
inbound logistics to the Bedford DC, for a specific men’s pant product which was sourced from a
contract manufacturer with its factory located in Nanjing, China near Shanghai for the 2011 season.
The systems boundary of this study is therefore set at the Shanghai Yangshan Deepwater
Port in China as the origin and REI’s eastern US distribution center in Bedford, Pennsylvania as the
final destination. The actual inbound transportation for the 2011 season pant product flowed from
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
3
the Port of Shanghai via ocean to the Port of Baltimore on the Atlantic coast, and from there via
road to the Bedford DC. Hence, this route will be used as the baseline where necessary.
Furthermore, the functional unit for the purpose of this study is assumed to be one full 40-foot
standard container load, or two twenty-foot equivalent units, of the pant product.
In order to keep this study focused and manageable, the following elements of the supply
chain are excluded from the scope:
Upstream supply chain beyond the origin port, due to infeasibility of data collection
Downstream supply chain beyond distribution centers (i.e., distribution center to retail
stores and customers) due to different characteristics of logistics
Inbound logistics to the western US distribution center, due to the high likelihood of
already achieving optimal state in terms of the objectives considered
Air freight, due to REI’s current practice of using air only for irregular, expedited
inbound shipments
2.2. Stage One
The goal of this stage is to identify intermodal freight transportation routes between the
Port of Shanghai and the Bedford DC that are available and suitable for analysis, through gaining a
broad understanding of international intermodal freight transportation. Information on intermodal
freight transportation in general, as well as on marine ports, ocean freight transportation, rail
freight transportation, and road freight transportation will be collected primarily through desktop
research, and supplemented by insights from practitioners engaged in transportation and logistics.
Types of information sources can be generally categorized as follows:
Government organizations
Environmental NGOs
Academia
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
4
Freight carriers and trade organizations
Ocean container ports and trade organizations
Logistics service providers
Shippers
Based on the information collected and understanding gained, the deliverable of this first
stage will be a narrowed-down list of inbound freight transportation modes and routes for data
collection and analysis in the subsequent stages.
2.3. Stage Two
The goal of this stage is to complete the list of inbound freight transportation routes
prepared in the previous stage by adding data for the objectives that will be considered in the
analysis in the following stage. The objectives in particular are total shipping costs, total transit
time and total GHG emissions per route. In order to calculate shipping costs, sub data such as
volume and mass of freight, distance of route, and price schedules will be collected. Similarly, in
order to calculate GHG emissions, sub data such as volume and mass of freight, distance of route,
and emission factors will be collected. Moreover, data will be required per transportation mode
that makes up each route.
Where possible, attempts will be made to collect primary data from the data owner. For
example, volume and mass of freight will be acquired from REI, and emission factors will be
acquired from the carrier, and so forth. In cases which primary data is not available, secondary data
will be obtained utilizing publicly available databases and tools. When neither primary nor
secondary data is available, proxy data and assumptions will be used.
Based on the data collected, the deliverable of this second stage will be a list of inbound
freight transportation routes with attributes in terms of shipping costs, transit time and greenhouse
emissions, for modeling and analysis in the following stage.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
5
2.4. Stage Three
The goal of this stage is to select the best intermodal route for inbound transportation from
the list created in the previous section based on the data collected, through applying a multiple
objective problem approach. A Microsoft Excel-based simple additive model of preferences will be
created to compute the overall utility of each route and to perform sensitivity analysis.
The simple additive model equation is:
, where
utility of route on attribute ,
utility of route on attribute ,
utility of route on attribute ,
weight assignment to attribute ,
weight assignment to attribute ,
weight assignment to attribute ,
Overall utility assigned to route .
Outcomes of attributes in the list from the previous section will be converted to utility on a
scale of 0 to 1 proportionately, with the best outcome of the attribute under consideration being a 1
and the worst being a 0. Assessment of tradeoff weights will use the pricing out method. Weight
assignments for the three attributes will be derived by setting the cost attribute as the numeraire
and determining the willingness to pay to go from the worst time to best time performance, and the
WTP to go from the worst emissions to best emissions performance. By substituting the converted
utilities and derived weight assignments into the simple additive model of preferences equation,
the overall utility for each route can be computed, thus revealing the best intermodal route for
inbound transportation.
Sensitivity analysis will be performed on the weight assignments. This will visualize how
the results are affected by preferences and the tradeoffs to be made.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
6
3. Findings and Results
In the first part of this section, the significance of freight transportation and inbound
logistics within the context of global warming is examined. In the remainder of the section, findings
and results are outlined according to the three stages of this study. Again, the first stage focused on
gaining an understanding of international intermodal transportation and identifying inbound
transportation modes and routes. The second stage was collecting input data to complete the list of
inbound transportation modes and routes with data on costs, transit time and emissions. The third
stage was modeling and analysis of the financial, operational and environmental impacts for
decision-making on the inbound logistics in question.
3.1. Freight Transportation in the Context of Global Warming
In the US in 2003, freight transportation sources accounted for approximately 438
teragrams of carbon dioxide-equivalents of GHG emissions, or 24.7 percent of GHG emissions from
all transportation sources and 6.3 percent of total GHG emissions (ICF Consultung, 2005). A similar
trend can be seen on the global level: A 2009 report estimates freight transportation accounts for
approximately 2,500 megatonnes of CO2e, or 5 percent of annual worldwide GHG emissions (World
Economic Forum, 2009). On both levels, road freight was the greatest contributor accounting for
77.8 percent of GHG emissions from freight transportation in the US and 63.8 percent of GHG
emissions from freight transportation worldwide.
Road freight being the biggest GHG emitter within the freight transportation sector does not
necessarily conclude it is the least environmentally-efficient; but it is the second least efficient
mode of transportation only after air freight. Figure 3-1 shows a comparison of emission factors by
freight transportation mode in terms of kilograms of CO2 emissions per ton-mile of freight moved,
adopted as default values for US vehicles in the GHG Protocol tool for mobile combustion (World
Resources Institute, 2012).
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
7
Figure 3-1: GHG Emission Factor by Mode of Freight Transportation (World Resources Institute, 2012)
In contrast to the data illustrated above, some studies and statistics point out that ocean
freight transportation is actually more environmentally-efficient than rail freight. For example,
ocean freight transportation is said to emit less than two-thirds of GHG per weight-distance
compared with rail (Dizikes, 2010), or ocean transportation is 32 to 55 percent more efficient than
rail at typical operating conditions (Herbert Engineering Corporation, 2011). This could be true
depending on which sets of data are used to derive an aggregated average. This study attempts to
address such issues by employing route- and mode-specific emission factors for analysis in the
following sections.
Finally, freight transportation and inbound logistics could be a particularly interesting area
to examine for apparel companies hoping to reduce their carbon footprint. In the case of Nike in
2009, inbound logistics accounted for 23 percent of the athletic apparel and footwear giant’s total
GHG footprint and was the second largest impact area only after manufacturing (Nike, Inc., 2010).
1.527
0.297
0.048
0.025
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
Air
Road
Ocean
Rail
kg CO2/ton-mile
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
8
3.2. Stage One
3.2.1. Intermodal Transportation
Gerhardt Muller (1999) defined intermodal freight transportation as “the concept of
transporting freight using more than one mode of travel in such a way that all parts of the
transportation process are effectively connected and coordinated, safe, environmentally sound, and
offering flexibility.” In practical terms, intermodal freight transportation allows moving goods over
multiple modes of transportation, in particular ocean, rail and road, without the handling of the
actual freight itself at the point of interchange.
In ocean freight, intermodal transportation is synonymous with transporting containerized
freight by container ships, as opposed to other types of freight transportation such as shipping oil
and chemical by tankers or ore and grain by bulk carriers. Generally, shipping containers, also
called intermodal containers or ISO containers, are 20 or 40 feet in length and 8 feet 6 inches in
height, hence container capacity is commonly expressed in twenty-foot equivalent units, or TEUs.
Although less common, 45-foot containers are also used in intermodal freight. 45-foot containers
are typically 9 feet 6 inches in height and this variant is called high cube. 40-foot containers come in
both standard and high cube variants. While these variations create a range in container volumes,
40-foot standard, 40-foot high cube and 45-foot high cube containers are all considered 2 TEU.
Rail freight adds more complexity; rail intermodal could be transporting containers on well
cars capable of double-stacking containers, containers on a flatcar or trailers on a flatcar. Further, in
the US, domestic containers are typically 48 or 53 feet long. These factors, among others,
necessitate a distinction between international and domestic intermodal freight transportation on
the rail carriers’ part. Union Pacific Railroad, a US Class I railroad, defines international intermodal
as “intact containerized US rail shipments involving an immediately prior or subsequent ocean
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
9
movement” (Union Pacific Railroad Company). As such, not all of the rail carriers’ intermodal
equipment and facilities are capable of handling international intermodal freight.
With regards to the state of intermodal transportation in the US, intermodal freight on rail
increased from 3 million containers and trailers in 1980 to 11.9 million units in 2011. Over the
same period, the share of containers rose from 42 percent to 85.6 percent (Association of American
Railroads, 2012). Double-stacking of containers was first introduced in the US in 1984, and by 2004,
accounted for approximately 70 percent of intermodal freight transportation on rail (Pacer
International, Inc., 2004).
3.2.2. Ports
Locations of the US top 25 container ports in 2009 in terms container traffic in TEU per year
are illustrated in Figure 3-2 (US Department of Transportation, 2011). The Port of New York/New
Jersey includes the Port of Newark. The Port of Norfolk is part of the Port of Virginia, which is
sometimes referred to as Hampton Roads. The Port of Los Angeles and the Port of Long Beach are
located next to each other much like the Port of New York/New Jersey, and sometimes treated as a
combined single port.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
10
Figure 3-2: US Top 25 Container Ports (US Department of Transportation, 2011)
Next, Table 3-1 lists North America’s top 25 container ports in 2010 in terms of container
traffic in TEU per year (American Association of Port Authorities). Combining the traffic of the Port
of Los Angeles and the Port of Long Beach would increase the number up to more than 14 million
TEU per year, nearly three times that of the Port of New York/New Jersey, the immediate follower.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
11
Rank Port Coast Country Container Traffic
(TEU/year) 1 Los Angeles Pacific Coast United States 7,831,902
2 Long Beach Pacific Coast United States 6,263,499
3 New York/New Jersey Atlantic Coast United States 5,292,025
4 Savannah Atlantic Coast United States 2,825,179
5 Metro Port Vancouver Pacific Coast Canada 2,514,309
6 Oakland Pacific Coast United States 2,330,214
7 Seattle Pacific Coast United States 2,133,548
8 Hampton Roads Atlantic Coast United States 1,895,017
9 Houston Gulf Coast United States 1,812,268
10 San Juan Atlantic Coast United States 1,525,532
11 Manzanillo Pacific Coast Mexico 1,509,378
12 Tacoma Pacific Coast United States 1,455,466
13 Charleston Atlantic Coast United States 1,364,504
14 Montreal Atlantic Coast Canada 1,331,351
15 Honolulu Pacific Coast United States 968,326
16 Jacksonville Atlantic Coast United States 857,374
17 Miami Atlantic Coast United States 847,249
18 Lazaro Cardenas Pacific Coast Mexico 796,011
19 Port Everglades Atlantic Coast United States 793,227
20 Veracruz Gulf Coast Mexico 677,596
21 Baltimore Atlantic Coast United States 610,922
22 Altamira Gulf Coast Mexico 488,013
23 Anchorage Pacific Coast United States 445,814
24 Halifax Atlantic Coast Canada 435,461
25 New Orleans Gulf Coast United States 427,518 Table 3-1: North America’s Top 25 Container Ports (American Association of Port Authorities)
Based on the findings on container ports in the US and North America, this study tentatively
narrowed down the routes for analysis to those that transit through a port in Table 3-1 and is
located on either the Pacific or Atlantic coast of contiguous US, with the exception of Metro Port
Vancouver. Routes for analysis are further examined in conjunction with the availability of ocean
and rail carriers at the ports.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
12
3.2.3. Ocean Freight
The world’s top 20 ocean container carriers in terms of operated capacity in TEU as of
January 2011 are presented in Table 3-2 (Cap-Marine Assurances & Réassurances SAS, 2011).
Maersk and Mediterranean Shipping Company are the two big players, with more than a 600,000
TEU capacity advantage ahead of CMA CGM Group, the number three.
Rank Carrier Operated Capacity (TEU)
Operated Capacity (Number of ships)
1 APM-Maersk 2,147,831 578 2 Mediterranean Shipping Company 1,863,449 450 3 CMA CGM Group 1,209,530 400 4 Evergreen Line 603,766 158 5 Hapag-Lloyd 596,774 136 6 APL 584,780 146 7 CSAV Group 579,296 155 8 COSCO Container Lines 544,857 139 9 Hanjin Shipping 476,955 104
10 China Shipping Container Lines 457,162 140 11 MOL Logistics 399,337 97 12 NYK Line 386,838 98 13 Hamburg Süd Group 370,851 116 14 OOCL 353,523 79 15 K Line 328,327 78 16 Zim Integrated Shipping Services 322,735 94 17 Yang Ming Marine Transport Corporation 322,091 79 18 Hyundai Merchant Marine 286,875 55 19 Pacific International Lines 263,558 142 20 United Arab Shipping Company 216,799 55
Table 3-2: World’s Top 20 Ocean Container Carriers (Cap-Marine Assurances & Réassurances SAS, 2011)
For the ocean freight segment, this study focused on Maersk, because the company’s Web
site provided richer information compared with that of other carriers, and further since
representatives from Maersk were willing to support this study by sharing data.
Maersk operates transpacific services from Shanghai to ports on both the Pacific and
Atlantic coasts of North America. Table 3-3 lists the Shanghai-originating services on transpacific
trade lanes as of October 2012 (Maersk Line).
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
13
Service Destination Port Destination Terminal Transit Days TP2 Long Beach, CA Total Terminals International/Pier T 15 TP3 Newark, NJ APM Terminal 32 TP3 Norfolk, VA APM Terminal 35 TP3 Savannah, GA Garden City Terminal 37 TP7 Miami, FL South Florida Container Terminal 25 TP7 Savannah, GA Garden City Terminal 26 TP7 Charleston, SC Wando Welch Terminal 28 TP8 Long Beach, CA Total Terminals International/Pier T 13 TP8 Oakland, CA International Container Terminal 17 TP9 Seattle, WA Terminal 18 13 TP9 Vancouver, Canada Deltaport Terminal 15
Table 3-3: Maersk Transpacific Services (Maersk Line)
Based on the findings on ocean freight carriers, this study narrowed down the routes for
analysis to those that go through the above landing ports called by Maersk. All of the above landing
ports were included in the tentative list from the screening conducted in the previous section. Ports
that were screened out from the tentative list were Tacoma, Jacksonville and Everglades.
Maersk was also a suitable ocean carrier to work with on this study, for its approaches to
environmental management and intermodal freight transportation. First, the carrier has been a
proponent of “slow steaming,” an operational strategy to slow vessels down which doesn’t require
adoption of new technology. According to Maersk, slow steaming helped the carrier reduce CO2
emissions per container by 12.5 percent over 2007 to 2009. A study by Cariou (2011) also
estimates that slow steaming led to an 11 percent CO2 emissions reduction in international shipping
over 2008 to 2010. Slow steaming not only cuts fuel consumption and thereby fuel costs and GHG
emissions, but also improves schedule reliability because it creates flexibility for ships to adjust
speed to meet delivery times (Maersk Line, 2010). In addition, in March 2012, Maersk introduced a
seamless intermodal freight service between major ports in Asia and Chicago, Dallas, Houston,
Memphis, and Northwest Ohio in the US via the Port of Los Angeles, in partnership with BNSF
Railway Company, a US Class I railroad. The collaboration between the ocean and rail carriers
allows faster transit and 95 percent on-time delivery (Maersk Line, 2012).
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
14
3.2.4. Rail Freight
There are five Class I railroads in the US providing freight transportation services, four of
which concern this study, namely BNSF Railway, Union Pacific Railroad, CSX Transportation, and
Norfolk Southern Railway. In addition, there are two Canadian Class I railways serving the
Vancouver area, namely Canadian National Railway and Canadian Pacific Railway.
3.2.4.1. BNSF Railway
BNSF mainly serves the western US region. Figure 3-3 presents the BNSF system map. The
Class I railroad has 25 intermodal facilities with international capability across its territory (BNSF
Railway Company, 2012). Of the landing ports which Maersk calls, BNSF has on-dock capability in
Seattle and Long Beach, and near-dock capability in Oakland. In Seattle, the Seattle International
Gateway – BNSF’s intermodal facility – is located just a half a mile from the port. At Long Beach,
BNSF’s Los Angeles Hobart intermodal facility is located approximately 20 miles from the port but
connected by the Alameda Corridor, a dedicated cargo rail expressway operated jointly by BNSF
and UP. In Oakland, although the Oakland International Gateway – also BNSF’s intermodal facility –
is a near-dock facility, it is located less than a mile away from the terminal.
Among the rail carriers that lack coverage in the eastern US region, BNSF is the only one
that publishes regular schedules for interline international intermodal services, implying a steel
wheel-based seamless interchange with eastern US railroads. From the Pacific coast to the Bedford
DC area, such routes are from the BNSF Los Angeles Hobart facility to CSXT facilities in Cleveland,
Ohio and Northwest Ohio, Ohio, and to NS facilities in Columbus, Ohio and Harrisburg, Pennsylvania.
International intermodal interchanges for other O-D pairs are also available, although they seem to
require individual arrangements.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
15
Figure 3-3: BNSF System Map (Wikipedia, 2009)
3.2.4.2. Union Pacific Railroad
UP mainly serves the western US region and competes with BNSF. Figure 3-4 is a
representation of the UP system map. The Class I railroad has 27 intermodal facilities with
international capability across its territory (Union Pacific Railroad Company, 2012). Of the landing
ports which Maersk calls, UP has on-dock capability in Seattle, and near-dock capability in Oakland
and Long Beach. UP’s intermodal facilities are located just a mile from the port in Seattle and on-
port in Oakland. At Long Beach, the Intermodal Container Transfer Facility – UP’s intermodal
facility serving both the Port of Los Angeles and the Port of Long Beach – is located 5 miles from
both ports, however, containers are drayed from the ports to the ICTF by truck.
In order to deliver international intermodal freight from the Pacific coast facilities to the
Bedford DC area, UP interchanges with CSXT and NS via the Chicago facilities, although such
interchanges seem to require individual arrangements. Within the UP system, international
intermodal service is available from Seattle to the Chicago Global II, Chicago Global III and Chicago
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
16
Global IV intermodal facilities; from Oakland to the Chicago Global IV intermodal facility; and from
the Long Beach ICTF to the Chicago Global III and Chicago Global IV intermodal facilities.
Figure 3-4: UP Systems Map (Wikipedia, 2009)
3.2.4.3. CSX Transportation
CSXT mainly serves the eastern US region. Figure 3-5 presents the CSXT system map. The
Class I railroad has 43 intermodal facilities with international capability across its territory (CSX
Transportation, Inc., 2012). Of the landing ports which Maersk calls, CSXT has near-dock capability
in Charleston, Savannah and Miami, as well as on-dock capability in Savannah. The near-dock
intermodal facilities are located approximately 14 miles from the port in Charleston, 7 miles from
the port in Savannah and 14 miles from the port in Miami, respectively, and containers are drayed
from the port to the intermodal facility by truck at all three locations. The Miami facility belongs to
and is operated by Florida East Coast Railway, an exclusive railroad for ports in South Florida which
interchanges with CSXT in Jacksonville, Florida. As of October 2012, a project to connect FEC to an
on-port rail facility at the Port of Miami is underway (Florida East Coast Railway, 2011).
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
17
In order to reach the Bedford DC area from South Atlantic ports, the CSXT intermodal
facility with international capability that allows the shortest dispatch via road is the one in
Baltimore, Maryland. Service is available to the Baltimore facility from the three near-dock facilities
in Charleston, Savannah and Miami, however, not from the on-dock facility in Savannah.
As previously discussed, CSXT provides seamless international intermodal interchange
from the BNSF Los Angeles Hobart facility to CSXT facilities in Cleveland and Northwest Ohio. CSXT
also provides steel wheel-based international intermodal interchange from BNSF and UP facilities
in Seattle, Oakland, Los Angeles, and Long Beach to CSXT facilities in Cleveland, Columbus and
Northwest Ohio via Chicago, although such interchanges seem to require individual arrangements.
Figure 3-5: CSXT System Map (Wikipedia, 2009)
3.2.4.4. Norfolk Southern Railway
NS mainly serves the eastern US region and competes with CSXT. Figure 3-6 is a
representation of the NS system map. The Class I railroad has 51 intermodal facilities with
international capability across its territory (Norfolk Southern Corp., 2012). Of the landing ports
which Maersk calls, NS also has near-dock capability in Charleston, Savannah and Miami, and on-
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
18
dock capability in Savannah. However, no international intermodal service is available from these
four facilities to NS facilities in Maryland or Pennsylvania, which provide similar or shorter access
compared with the CSXT Baltimore facility to the Bedford DC.
As previously discussed, NS provides seamless international intermodal interchange from
the BNSF Los Angeles Hobart facility to NS facilities in Columbus and Harrisburg. Information on
availability and schedules of other international intermodal interchanges from western US
railroads to the Bedford DC area could not be obtained.
Figure 3-6: NS System Map (Wikipedia, 2009)
3.2.4.5. Canadian National Railway
CN mainly serves Canada and parts of US. Figure 3-7 presents the CN system map. Of the
landing ports which Maersk calls, CN has on-dock capability in Vancouver (Canadian National
Railway Company, 2012). The CN Vancouver intermodal facility is located in Surrey, British
Columbia, approximately 28 miles from the port. International intermodal service is available from
Vancouver to the CN Chicago facility in Harvey, Illinois. An agreement with CSXT to provide steel
wheel-based interchange in Chicago was announced in April 2012, however, information on start of
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
19
service and schedules could not be obtained. Prior to this agreement, CN and CSXT exchanged
container traffic in Chicago only by truck (Canadian National Railway Company, 2012).
Figure 3-7: CN System Map (Wikipedia, 2009)
3.2.4.6. Canadian Pacific Railway
CPR mainly serves Canada and parts of US and competes with CN. Figure 3-8 is a
representation of the CPR system map. Of the landing ports which Maersk calls, CPR has on-dock
capability in Vancouver (Canadian Pacific, 2012). The CPR Vancouver intermodal facility is located
in Pitt Meadows, British Columbia, approximately 35 miles from the port. International intermodal
service is available from Vancouver to the CPR Chicago facility in Bensenville, Illinois, however,
information on availability and schedules of interchanges to eastern US railroads could not be
obtained.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
20
Figure 3-8: CPR System Map (Wikipedia, 2009)
3.2.5. Road Freight
The destination terminals from where the international intermodal freight is dispatched via
road to the Bedford DC have been identified from findings in the previous sections. Table 3-4
summarizes the terminals, their location and distance to the Bedford DC.
Terminal Location Road Miles NS Harrisburg 3500 Industrial Rd., Harrisburg, PA 17110 104 CSXT Baltimore 4801 Keith Ave., Baltimore, MD, 21224 148 Port of Baltimore 2700 Broening Hwy., Baltimore, MD 21224 149 CSXT Cleveland 601 E 152nd St., Cleveland, OH 44110 239 Port of Newark 5080 McLester St., Elizabeth, NJ 07207 262 CSXT Columbus 2351 Westbelt Dr., Columbus, OH 43228 285 NS Columbus-Rickenbacker 3329 Thoroughbred Dr., Lockbourne, OH 43217 289 CSXT Northwest Ohio 17000 Deshler Rd., North Baltimore, OH 45872 357
Table 3-4: Terminal Locations and Distance to Bedford Distribution Center
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
21
Figure 3-9 is a map of the selected terminals relative to the Bedford DC.
Figure 3-9: Map of Terminal Locations and Bedford Distribution Center
3.2.6. Routes and Modes of Transportation for Data Collection
Based on the findings from the previous sections, 14 combinations of alternative routes and
modes of inbound transportation in addition to the baseline were defined for data collection. The
routes can be broadly-divided into four groups based on the location of their landing port. In
particular, Routes 1 through 3 can be grouped as the Pacific Northwest group, Routes 4 through 10
the California group, Routes 11 through 13 the South Atlantic group, and Routes 14 and 15 the Mid
Atlantic group. All groups but the Mid Atlantic use rail from the landing port to one of the
intermodal rail terminals within 360 miles from the Bedford DC. The remaining segments of the
routes use truck. The 15 routes are listed in Table 3-5.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
22
Origin Destination Mode Carrier Route 1 Port of Shanghai Port of Seattle Ocean Maersk BNSF Seattle BNSF Chicago Rail BNSF BNSF Chicago CSXT Cleveland Rail CSXT CSXT Cleveland Bedford DC Road N/A Route 2 Port of Shanghai Port of Seattle Ocean Maersk BNSF Seattle BNSF Chicago Rail BNSF BNSF Chicago CSXT Columbus Rail CSXT CSXT Columbus Bedford DC Road N/A Route 3 Port of Shanghai Port of Seattle Ocean Maersk BNSF Seattle BNSF Chicago Rail BNSF BNSF Chicago CSXT Northwest Ohio Rail CSXT CSXT Northwest Ohio Bedford DC Road N/A Route 4 Port of Shanghai Port of Oakland Ocean Maersk BNSF Oakland BNSF Chicago Rail BNSF BNSF Chicago CSXT Cleveland Rail CSXT CSXT Cleveland Bedford DC Road N/A Route 5 Port of Shanghai Port of Oakland Ocean Maersk BNSF Oakland BNSF Chicago Rail BNSF BNSF Chicago CSXT Columbus Rail CSXT CSXT Columbus Bedford DC Road N/A Route 6 Port of Shanghai Port of Oakland Ocean Maersk BNSF Oakland BNSF Chicago Rail BNSF BNSF Chicago CSXT Northwest Ohio Rail CSXT CSXT Northwest Ohio Bedford DC Road N/A Route 7 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart NS Harrisburg Rail BNSF-NS NS Harrisburg Bedford DC Road N/A Route 8 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart CSXT Cleveland Rail BNSF-CSXT CSXT Cleveland Bedford DC Road N/A Route 9 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart NS Columbus-Rickenbacker Rail BNSF-NS NS Columbus-Rickenbacker Bedford DC Road N/A
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
23
Route 10 Port of Shanghai Port of Long Beach Ocean Maersk BNSF Long Beach BNSF Los Angeles Hobart Rail BNSF BNSF Los Angeles Hobart CSXT Northwest Ohio Rail BNSF-CSXT CSXT Northwest Ohio Bedford DC Road N/A Route 11 Port of Shanghai Port of Miami Ocean Maersk Port of Miami FEC Miami Road N/A FEC Miami CSXT Baltimore Rail FEC-CSXT CSXT Baltimore Bedford DC Road N/A Route 12 Port of Shanghai Port of Savannah Ocean Maersk Port of Savannah CSXT Savannah Road N/A CSXT Savannah CSXT Baltimore Rail CSXT CSXT Baltimore Bedford DC Road N/A Route 13 Port of Shanghai Port of Charleston Ocean Maersk Port of Charleston CSXT Charleston Road N/A CSXT Charleston CSXT Baltimore Rail CSXT CSXT Baltimore Bedford DC Road N/A Route 14 Port of Shanghai Port of Newark Ocean Maersk Port of Newark Bedford DC Road N/A Route 15 (baseline) Port of Shanghai Port of Baltimore Ocean Maersk Port of Baltimore Bedford DC Road N/A
Table 3-5: Routes and Modes of Transportation for Data Collection
Routes that have been eliminated are:
UP routes, due to uncertain availability of interchange to eastern US railroads, and
overlap with BNSF routes
Port of Vancouver routes, due to uncertain availability of interchange from CN and CPR
to eastern US railroads
Port of Norfolk route, due to long ocean transit time and long road distance to Bedford
3.3. Stage Two
This stage focused on information and data collection to expand the list of routes and modes
of transportation from the previous stage for detailed analysis. The following data were collected:
Volume and mass
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
24
Distance
Costs
Transit time
Emission factors
Due to unavailability of sufficient data, the following components of inbound transportation
were excluded from this study:
Activities at ocean ports, such as loading, unloading and customs clearing
Activities at rail yards and terminals, such as switching, loading and unloading
3.3.1. Volume and Mass
Based on information and data provided by REI, the pant product is packed in various
quantities into cartons, loose-loaded without palletizing onto containers and transported together
with other products to the US over multiple shipments. The product weighs 13 ounces and 40 pairs
can be packed into a carton at maximum. Cartons are 0.60 meters long, 0.40 meters wide and 0.35
meters tall, and weigh 1.1 kilograms.
Internal dimensions of a typical 40-foot standard container are 12.03 meters in length, 2.35
meters in width and 2.39 meters in height. Tare weight of the container is 3,700 kilograms (Maersk
Line). For the purpose of this study, it is assumed that such a 40-foot standard container is loaded
to maximum capacity with cartons of only the pant product. It is also assumed that no pallets are
used. Based on an optimal stowage pattern of cartons on the container floor, 110 cartons can be laid
out per tier and stacked up six tiers. The payload weight, consisting of 660 cartons each packed
with 40 products, is 10,456 kilograms. Adding the tare weight of the container, the gross weight
becomes 14,156 kilograms, or 15.6 tons. Therefore, the functional unit of this study is one 40-foot
standard container, or 2 TEU, or 15.6 tons.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
25
3.3.2. Distance
Ocean distance data was acquired primarily from Maersk. Where additional data was
required, materials from the National Oceanic and Atmospheric Administration (US Department of
Commerce, 2009) and the PortWorld Web site (Petromedia Ltd., 2012) were used.
Rail and road distance data were collected using PC*MILER, a routing, mileage and mapping
software for the transportation and logistics industry developed by ALK Technologies. In particular,
the PC*MILER Rail software hosts data for various modes of rail transport including intermodal
freight, and is adopted by Class I railroads such as BNSF, UP and CSXT for price calculation tools
based on rail distances. Trial versions of PC*MILER Version 26 and PC*MILER Rail Version 18 were
used for road and rail, respectively.
Distances in terms of miles by transportation mode per route are summarized in Table 3-6.
Route Ocean Rail Road Total Distance
1 5,851 2,734 239 8,825 2 5,851 2,756 285 8,893 3 5,851 2,706 357 8,915 4 6,971 2,846 239 10,056 5 6,971 2,868 285 10,124 6 6,971 2,818 357 10,146 7 6,537 3,308 104 9,949 8 6,537 2,555 239 9,332 9 6,537 2,596 289 9,423
10 6,537 2,524 357 9,418 11 11,210 1,162 161 12,533 12 11,676 647 154 12,477 13 11,793 545 162 12,500 14 12,221 0 262 12,483 15 12,338 0 149 12,487
Table 3-6: Distance per Route by Mode of Transportation
3.3.3. Emission Factors
Ocean emission factors were acquired from Maersk which are service-specific, based mostly
on data from 2011 and partially 2010. While emission factors for freight transportation are
generally expressed in emissions per weight-distance units (e.g., kilogram CO2 per tonne-kilometer),
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
26
those for ocean container transportation are typically expressed in emissions per volume-distance
units (e.g., kilogram CO2 per TEU-kilometer), and Maersk follows this practice.
Compared with WRI’s GHG Protocol tool for mobile combustion previously discussed,
Maersk’s emission factors, when converted to weight-distance unit, are much smaller than the
default value for ocean freight set in the tool, in fact, smaller than that of rail freight as some suggest.
Some factors that affect fuel efficiency and thus carbon emissions are speed, load, vessel and engine
type and age, and capacity utilization (Herbert Engineering Corporation, 2011). Given that Maersk
has been practicing slow steaming, it could be contributing to the variance between the emission
factors of Maersk and the GHG Protocol tool, together with the data aggregation issues previously
discussed. In addition, the inexactness of TEU could also be a source of variance when converting
volume-distance emission factors to weight-distance emission factors, which could be another
cause of the different opinions regarding carbon efficiencies of ocean and rail transportation.
Emission factors for rail and road were obtained from the US EPA SmartWay Carrier Data
(US Environmental Protection Agency, 2012). For rail, intermodal-specific emission factors by
railroad were available. For road, average of 103 carriers’ data for dray and average of 13 carriers’
data for truck intermodal, respectively, were adopted for the purpose of this study.
Emission factors by mode of transportation are summarized in Table 3-7.
Ocean
(kg CO2/TEU-mile)
Rail
(kg CO2/ton-mile)
Road Intermodal
(kg CO2/ton-mile)
Road Dray
(kg CO2/ton-mile) 0.0829-0.0898 0.0200-0.0242 0.0531 0.0866
Table 3-7: Emission Factors by Mode of Transportation
3.3.4. Costs
Ocean freight rates as of November 2012 were collected from the rate search tool on
Maersk’s Web site (Maersk Line). Maersk’s rates are charged per container and different rates
apply for different types of container. Road freight costs were taken from the PC*MILER tool.
Default settings of PC*MILER were used, of which the major components are fuel, toll and driver
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
27
costs. It was assumed that these costs are charged per container and that the same costs apply
regardless of container type.
Rail intermodal prices seem to generally consist of two components, a station-to-station
price and a fuel surcharge. However, none of the rail carriers but CN publishes their intermodal
prices. For the purpose of this study, CN intermodal prices were taken as a proxy and applied to the
15 routes for analysis. For the station-to-station component, prices of 50 sample CN O-D pairs
(Canadian National Railway Company, 2012) were converted to an average price per mile of $1.12.
CN’s fuel surcharge is 17.37 percent of the station-to-station price, as of November 2012 (Canadian
National Railway Company, 2005). CN prices are per container and the same price applies for 20-
foot, 40-foot and 45-foot containers. For the Alameda Corridor segment of the Port of Long Beach
routes, the published use fee of $21.60 per TEU was used (Alameda Corridor Transportation
Authority, 2012).
Costs in US dollars for one 40-foot standard container by transportation mode are
summarized in Table 3-8.
Route Ocean Rail Road Total Costs 1 3,907 3,594 401 7,902 2 3,907 3,623 418 7,949 3 3,907 3,558 564 8,029 4 4,007 3,741 401 8,149 5 4,007 3,770 418 8,196 6 4,007 3,705 564 8,276 7 4,007 4,363 177 8,547 8 4,007 3,374 401 7,782 9 4,007 3,428 428 7,862
10 4,007 3,332 564 7,903 11 5,557 1,528 252 7,337 12 5,525 851 240 6,616 13 5,525 717 250 6,492 14 5,585 0 414 5,999 15 6,341 0 232 6,573
Table 3-8: Total Costs per Route by Mode of Transportation
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
28
3.3.5. Transit Time
Transit times for ocean and rail freight were acquired from schedules published on carriers’
Web sites (Maersk Line) (BNSF Railway Company, 2012) (CSX Transportation, Inc., 2012). When
two or more transit times were quoted for a single service they were averaged. Concerning rail, one
additional day of transit time was added to routes that have non-seamless interchange, namely
Routes 1 through 3, the Port of Seattle routes, and Routes 4 through 6, the Port of Oakland routes.
An assumption of 45 minutes was made for the Alameda Corridor segment of Routes 7 through 10,
namely the Port of Long Beach routes, based on the average speed of trains of 35 to 40 miles per
hour over the 21.6-mile segment (Net Resources International). Transit times for road were
obtained from PC*MILER.
Transit times by transportation mode in terms of days, hours and minutes are presented in
Table 3-9 in d:hh:mm format.
Route Ocean Rail Road Total Transit Time
1 13:12:00 9:12:20 0:03:48 23:04:08 2 13:12:00 9:18:12 0:04:31 23:10:43 3 13:12:00 9:05:42 0:05:51 22:23:33 4 16:00:00 8:17:08 0:03:48 24:20:56 5 16:00:00 8:23:00 0:04:31 25:03:31 6 16:00:00 8:10:30 0:05:51 24:16:21 7 12:00:00 6:23:45 0:01:43 19:01:28 8 12:00:00 7:14:29 0:03:48 19:18:17 9 12:00:00 6:17:45 0:04:41 18:22:26
10 12:00:00 1:07:15 0:05:51 13:13:06 11 25:00:00 3:06:21 0:02:43 28:09:04 12 27:12:00 1:05:42 0:02:33 28:20:15 13 28:12:00 0:23:42 0:02:40 29:14:22 14 32:00:00 0:00:00 0:04:14 32:04:14 15 41:12:00 0:00:00 0:02:26 41:14:26
Table 3-9: Transit Times per Route by Mode of Transportation
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
29
3.3.6. Greenhouse Gas Emissions
GHG emissions were calculated by summing the products of volume or weight by distance
by emission factor per transportation mode. Total emissions in terms of kg CO2 by transportation
mode are summarized in Table 3-10.
Route Ocean Rail Road Total Emissions
1 1,051 861 198 2,110 2 1,051 868 236 2,154 3 1,051 852 296 2,199 4 1,156 896 198 2,250 5 1,156 903 236 2,294 6 1,156 887 296 2,338 7 1,084 1,111 86 2,281 8 1,084 804 198 2,086 9 1,084 842 239 2,165
10 1,084 792 296 2,171 11 2,013 367 141 2,521 12 2,097 202 131 2,430 13 2,118 170 141 2,429 14 2,195 0 217 2,412 15 2,216 0 124 2,339
Table 3-10: Total Greenhouse Gas Emissions per Route by Mode of Transportation
3.3.7. Routes and Modes of Transportation for Analysis
With the data collected, the list of routes and modes of transportation for analysis was
completed. The list is provided in Table 3-11.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
30
Route Costs
(US$)
Transit Time
(d:hh:mm)
GHG Emissions (kg CO2)
1 7,902 23:04:08 2,110 2 7,949 23:10:43 2,154 3 8,029 22:23:33 2,199 4 8,149 24:20:56 2,250 5 8,196 25:03:31 2,294 6 8,276 24:16:21 2,338 7 8,547 19:01:28 2,281 8 7,782 19:18:17 2,086 9 7,862 18:22:26 2,165
10 7,903 13:13:06 2,171 11 7,337 28:09:04 2,521 12 6,616 28:20:15 2,430 13 6,492 29:14:22 2,429 14 5,999 32:04:14 2,412 15 6,573 41:14:26 2,339
Table 3-11: Routes and Modes of Transportation for Analysis
3.4. Stage Three
The major purpose of this study was to provide a practical framework for decision-making
in an environment with multiple objectives involving tradeoffs. Given the 15 alternatives to choose
from, how should the logistics manager decide on which route to ship his pant product from
Shanghai to the Bedford DC? The following sections attempt to address this question.
3.4.1. Multiple Objective Decision-Making
The objectives of this decision-making question on inbound transportation route selection
are to keep the three attributes of costs, transit time and emissions all low. The study applied a
simple additive model of preferences to address this multiple objective problem. As previously
discussed, the goal was to compute the overall utility of each route using the following equation:
Procedures to reach this computation are presented in this section step-by-step.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
31
3.4.1.1. Ranking and Dominance Check
First, outcomes of each attribute were converted to rankings as listed in Table 3-12.
Route Costs Transit Time GHG Emissions
1 8 6 2 2 10 7 3 3 11 5 6 4 12 9 7 5 13 10 9 6 14 8 10 7 15 3 8 8 6 4 1 9 7 2 4
10 9 1 5 11 5 11 15 12 4 12 14 13 2 13 13 14 1 14 12 15 3 15 11
Table 3-12: Ranking of Attributes
Figure 3-10 illustrates the attribute rankings in a chart format. It can be observed that
Route 8 dominates Routes 1 through 6. At this point, the three Port of Seattle routes and the three
Port of Oakland routes can be eliminated from further consideration, since the Long Beach to CSXT
Cleveland route is superior on all three attributes. However, dominance does not hold for the
remaining nine alternatives. It is visually and theoretically confirmed that there are tradeoffs
among the three objectives.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
32
Figure 3-10: Ranking of Attributes
3.4.1.2. Trade-Off Weight Assessment
Next, a range of best and worst levels were defined for each attribute. The levels could
simply be set at the best and worst outcomes of the alternatives under consideration, or chosen
otherwise. For the purpose of this study, the former was employed as described in Table 3-13.
Costs
(USD)
Transit Time
(d:hh:mm)
GHG Emissions (kg CO2)
Best 5,999 13:13:06 2,086 Worst 8,547 41:14:26 2,521
Table 3-13: Range of Attributes
Using the defined ranges, outcomes of attributes described in Table 3-11 were converted to
utilities on a scale of 0 to 1 proportionately, with the best outcome of the attribute under
consideration being a 1 and the worst being a 0. The converted utilities are listed in Table 3-14.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Cost Time Emissions
Ran
k
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
33
Route Costs Transit Time GHG Emissions
1 0.25 0.66 0.95 2 0.23 0.65 0.84 3 0.20 0.66 0.74 4 0.16 0.60 0.62 5 0.14 0.59 0.52 6 0.11 0.60 0.42 7 0.00 0.80 0.55 8 0.30 0.78 1.00 9 0.27 0.81 0.82
10 0.25 1.00 0.81 11 0.47 0.47 0.00 12 0.76 0.45 0.21 13 0.81 0.43 0.21 14 1.00 0.34 0.25 15 0.77 0.00 0.42
Table 3-14: Utility Scores of Attributes
The next step was to assess the tradeoff weights using the pricing out method. First, the
logistics manager needs to determine the maximum cost he is willing to pay if he were to improve
his transit time from worst performance to best performance. For the purpose of this study, it was
arbitrarily assumed that he is willing to pay the average cost of all the outcomes under
consideration which is $7,574. Conceptually, this is saying that if the cost of the expensive, fastest
route is at this level, the logistics manager would be indifferent between that route and the
inexpensive, slowest route. In mathematical terms, it is expressed as:
( ) ( ) ( ) ( )
The WTP of $7,574 is converted to a utility score of 0.38 using the cost utility curve from the
attribute range. By substituting utilities, the above equation becomes:
And by solving for , the equation becomes:
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
34
The same is done for the emissions attribute. For the purpose of this study, it was arbitrarily
assumed that the logistics manager is willing to pay the second highest cost of all the outcomes
under consideration which is $8,276, to improve his emissions performance from worst to best.
( ) ( ) ( ) ( )
Utility for WTP of $8,276 is 0.11 from the cost utility curve.
And by solving for , the equation becomes:
Finally, the sum of the three attribute weights must equal 1. Therefore:
The weights for the three attributes can be derived:
⁄
( )( )
( )( )
With the weights derived per above, the overall utility for each route can be computed.
Results are presented in Table 3-15.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
35
Route Costs Transit Time GHG Emissions
Overall Utility Overall Rank
1 0.14 0.15 0.20 0.49 8 2 0.13 0.14 0.18 0.45 9 3 0.12 0.15 0.15 0.42 10 4 0.09 0.13 0.13 0.35 12 5 0.08 0.13 0.11 0.32 13 6 0.06 0.13 0.09 0.28 15 7 0.00 0.18 0.12 0.29 14 8 0.17 0.17 0.21 0.55 4 9 0.15 0.18 0.17 0.50 7
10 0.14 0.22 0.17 0.53 5 11 0.27 0.11 0.00 0.37 11 12 0.43 0.10 0.04 0.58 3 13 0.46 0.10 0.04 0.60 2 14 0.57 0.07 0.05 0.70 1 15 0.44 0.00 0.09 0.53 6
Table 3-15: Weighted and Overall Utility Scores and Ranking
Route 14, the Port of Newark route, scored the highest overall utility. Therefore, this is
logically the best route to ship one full 40-foot standard container load of the pant product from the
Port of Shanghai to the Bedford DC, based on the ranges of the attributes that were defined and the
amounts of WTP that were determined.
3.4.2. Sensitivity Analysis
Two types of sensitivity analysis were performed. First, how the best route is affected by
varying the weights of attributes was tested. Next, the assumption made for the uncertain station-
to-station price for rail was examined.
3.4.2.1. Sensitivity Analysis of Weighting Assessment
Sensitivity analysis of the weighting assessment was performed by increasing the weight of
one attribute while holding all other variables constant. Figure 3-11 illustrates how the overall
utility scores vary as the weight on the emissions attribute is increased from the base case weight of
0.209. It is observed that Route 8, the Long Beach to CSXT Cleveland route, becomes the preferred
alternative as the weight on emissions is increased.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
36
Figure 3-11: Sensitivity Analysis by Increasing Weight of Emissions Attribute
The same was done on the transit time attribute. Route 10, the Long Beach to CSXT
Northwest Ohio route becomes the preferred alternative as the weight on transit time is increased
from the base case weight of 0.223. Figure 3-12 illustrates the results.
Figure 3-12: Sensitivity Analysis by Increasing Weight of Time Attribute
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Emissions
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Time
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
37
3.4.2.2. Sensitivity Analysis of Assumptions
The most uncertain data used in this study was the station-to-station price for rail. As
previously discussed, since none of BNSF, CSXT and NS publishes intermodal prices, those of CN, a
Canadian rail carrier, were averaged and used as a proxy. Sensitivity analysis of the cost per mile of
rail was performed by decreasing the assumption of $1.12 per mile while holding all other variables
constant. The dominance check conducted in Section 3.4.1.1 is not applicable (i.e., all 15 alternatives
need to be included) for this analysis, since varying this assumption will alter the original outcomes
and potentially the rankings of the Cost attribute.
Figure 3-13 illustrates how the overall utility scores vary as the cost per mile of rail is
decreased. It is observed that Route 8, the Long Beach to CSXT Cleveland route, could become the
best route as the cost per mile of rail is decreased.
Figure 3-13: Sensitivity Analysis by Decreasing Cost per Mile of Rail Transportation
Holding the weight assignments derived in the base case constant, the indifference point
between Route 14 and Route 8 is calculated to be at approximately $0.91 per mile. In other words,
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Ove
rall
uti
lity
Rail point-to-point price (US$/mi)
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
38
in the base case, the best route determined by this analysis is not impacted by the inaccuracy of the
rail cost as long as the true cost per mile of rail is not less than $0.91.
3.4.3. Other Types of Containers
In addition to the functional unit of this study of one 40-foot standard container, analyses
were performed for one 20-foot standard container, one 40-foot high cube container and one 45-
foot high cube container. Comparison of the results along with volume and mass of the respective
container types are provided in Table 3-16. The same assumptions for the price out assessment
were made for all types of containers.
20’ standard 40’ standard 40’ high cube 45’ high cube Number of containers 1 1 1 1 Volume (TEU) 1 2 2 2 Gross weight (ton) 7.9 15.6 17.7 20.2 Best route Route 14 Route 14 Route 14 Route 14 Potentially best route when weight on Emissions attribute is increased
Route 8 Route 8 Route 8 Route 14
Potentially best route when weight on Time attribute is increased
Route 10 Route 10 Route 10 Route 10
Potentially best route when Cost per mile of rail assumption is decreased
Route 8 Route 8 Route 8 Route 8 then Route 1 then Route 7
Table 3-16: Comparison of Results by Container Type
While the best route turned out to be the same for all container types under the
assumptions made, sensitivity analysis for 45-foot high cube container produced different results
compared with other types of containers.
4. Discussion of Findings and Future Research
To ship one full 40-foot standard container load of the pant product from the Port of
Shanghai to the Bedford DC, the ranking and dominance check first revealed that Routes 1 through
6, the Port of Seattle routes and the Port of Oakland routes, can be eliminated from further
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
39
consideration since they are dominated by Route 8, the Port of Long Beach to Cleveland route. Next,
the simple additive model of preferences concluded that Route 14, the Port of Newark route is the
best route, given the ranges of attributes defined and the WTP determined from pricing out.
Further, the sensitivity analysis on weight assignments presented that Route 8 could become the
preferred alternative as the weight on emissions is increased, and that Route 10, the Port of Long
Beach to Northwest Ohio route, could become the preferred alternative as the weight on transit
time is increased. Finally, the sensitivity analysis on cost per mile of rail indicated that given the
weight assignments of the base case, the result of Route 14 being the best route holds unless the
true cost per mile of rail is less than $0.91 versus the assumption of $1.12. In case the true cost per
mile of rail is less than $0.91, Route 8 becomes the best route.
Results of analysis for a particular container type cannot be generally applied across
different types of containers. This is due primarily to the different denominators of emission factors
and price schedules among different modes of transportation. For example, because emission
factors for ocean freight transportation are provided in per TEU-distance unit, emissions for the
ocean segment of a single route is the same between a 40-foot standard container and a 45-foot
high cube container, which are both 2 TEU. However, because emission factors for rail freight
transportation are provided in per weight-distance unit, emissions for the rail segment of the same
route are different between 40-foot standard and 45-foot high cube containers, which in this study,
the gross weight were assumed to be 15.6 tons and 20.2 tons, respectively. These could cause the
overall utility of a route with higher mix of rail to degrade relative to routes with lower mix of or no
rail as the gross weight of freight increases. Therefore, multiple objective analysis and tradeoff
weight assessment should be conducted based on the specific type of container used.
While accurate data for rail costs and a rigorous approach for determining the WTP would
make this analysis theoretically more robust, one of the objectives of this study was to provide the
logistics manager with a practical aid for decision-making on the selection of his preferred inbound
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
40
transportation route. Based on this analysis the logistics manager is already in a better position to
make an informed decision by gaining an understanding of how his choice is impacted by his
preferences, the tradeoffs he is willing to make and uncertain data. In any case, decision-making
should not rely solely on data or models, and in that respect, the model used in this study should
serve as an aid for decision-making rather than an automatic route selection tool.
Nevertheless, there are shortcomings of this model that can be improved. In terms of
accuracy, in addition to the uncertain data for rail costs, the exclusion of activities at marine ports
and rail facilities is another uncertainty that could potentially alter the outcomes when addressed.
In terms of scale, the model was built to analyze only one specific O-D pair of inbound
transportation; adding data for multiple O-D pairs would add more use to the model. In so doing,
adding overseas origin points to the model is relatively simple. However, increasing the number of
domestic destinations will likely require considerable efforts due to the complexity of interline
interchanging associated with rail transportation. Furthermore, enhancing the scope to include
outbound logistics would be another research area of its own, due to the different characteristics
and needs compared with inbound logistics.
Such improvements to the model or a development of a more versatile, generic platform
could be areas of interest for future research. Whichever direction that might take, one common
key to success will be cross-sector collaboration. By nature, intermodal freight transportation
involves many different players, and typically, there is lack of transparency at some point along the
supply chain, especially at points of interchange. The BSR Clean Cargo Working Group has been
working to bring participants primarily from ocean freight transportation together, and the US EPA
SmartWay program similarly to bring participants from domestic road and rail freight
transportation together. Although the main focus of both organizations is to address environmental
impacts, they could be potential partners for pursuing future research effectively and efficiently.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
41
5. Conclusion
Logistics managers confront tradeoffs among keeping costs low, moving goods promptly
and reducing carbon footprint. Multiple objective analysis can be one practical framework to help
decision-making on problems involving such tradeoffs at the intersection of business and the
environment. This study attempted to demonstrate such an application by addressing a real-world
problem REI faces regarding the choice of intermodal freight transportation routes for its inbound
logistics. The application of the framework revealed, quantitatively and visually, how the logistics
manager’s choice is affected by his preferences and the tradeoffs he is willing to make.
The methodologies employed for this study were fairly straight forward; however,
collection of input data presented a challenge. In particular, there is limited information and data
publicly available regarding the rail segment of international intermodal freight transportation,
especially when interline interchange is involved. This could be due to the fact that evolution is
currently ongoing in this field. Carriers are investing in equipment and infrastructure and exploring
partnerships to make end-to-end freight transportation more seamless and efficient for shippers.
As evidenced by the launch of the seamless intermodal freight service by Maersk in partnership
with BNSF, standardized one-stop-shopping and one-stop-billing services could become the norm
as this evolution unfolds. Such trends should also drive improvements in data consistency,
transparency and accessibility for shippers.
Nevertheless, the current limitation in availability of data especially for the rail segment of
international intermodal freight transportation could hinder informed decision-making by shippers
to optimize shipping costs, transit time and carbon footprint of their inbound logistics. On the flip
side, it could also mean missed business opportunity for the rail carriers; given the assumptions,
the result of the base case of this study suggests the preferred intermodal route from the Port of
Shanghai to the Bedford DC is to not go via rail. What if the use of true costs in the model altered
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
42
this outcome? Furthermore, as indicated by the example of Nike, the impact of inbound logistics
within the shipper’s corporate value chain could be one not to be neglected.
A versatile platform loaded and maintained with accurate and consistent cost, transit time
and emissions data, covering multipoint-to-multipoint international intermodal freight
transportation routes, could be a valuable aid for decision-making by shippers to enhance business
and environmental performance. Cross-sector forums on freight transportation such as the BSR
Clean Cargo Working Group and the US EPA SmartWay program could be leveraged to gauge
interest in and facilitate development of such a platform. The Duke Center for Sustainability and
Commerce is uniquely positioned to drive such a pursuit with its expertise in corporate
sustainability, network with players from industry, government and NGOs, and resources offered
by the greater Duke community.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
43
References
Alameda Corridor Transportation Authority. (2012, January 1). 2012 Charge per TEU (Twenty Foot Equivalent Unit). Retrieved from Alameda Corridor Transportation Authority Web site: http://www.acta.org/gen/charge_per_teu_2012.pdf
American Association of Port Authorities. (n.d.). North American Container Port Traffic (1990-2010). Retrieved from American Association of Port Authorities Web site: http://www.aapa-ports.org/Industry/content.cfm?ItemNumber=900#Statistics
Ashar, A., & Swigart, S. (2007). A Comparative Analysis of Intermodal Ship-to-Rail Connections at Louisiana Deep Water Ports. Retrieved from Louisiana Department of Transportation and Development Web site: http://www.dotd.la.gov/intermodal/marineandrail/documents/A_Comparative_Analysis_of_Intermodal_Ship_to_Rail_Connections_at_Louisiana_Deep_Water_Ports.pdf
Association of American Railroads. (2012, July). Rail Intermodal Keeps America Moving. Retrieved from Association of American Railroads Web site: http://www.aar.org/~/media/aar/Background-Papers/Rail-Intermodal.ashx
BNSF Railway Company. (2012). BNSF Railway Carbon Estimator. Retrieved from BNSF Railway Company Web site: http://domino.bnsf.com/website/carbon.nsf/ce?open
BNSF Railway Company. (2012, October 19). Intermodal Schedules. Retrieved from BNSF Railway Company Web site: http://www.bnsf.com/customers/pdf/international_P.pdf
BSR. (2009, October 19). Clean Cargo Working Group Intermodal Tool Training. BSR. (2011, March). How Brands Can Improve Supply Chain Sustainability Through the Clean Cargo
Working Group. Retrieved from BSR Web site: http://www.bsr.org/reports/CCWG_Report_Mar_2011_FINAL.pdf
Canadian National Railway Company. (2005, April 1). Fuel Surcharge. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/shipping-prices-tariffs-fuel-surcharge.htm
Canadian National Railway Company. (2012). CN Intermodal Terminals List. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/shipping-how-intermodal-terminals.htm
Canadian National Railway Company. (2012, April 2). CN offers new steel-wheel-interchange service with CSXT in Chicago, giving west coast imports efficient access to Ohio Valley markets. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/media-news-interchange-service-csx-20120402.htm
Canadian National Railway Company. (2012). CN T 008383-BQ INTERMODAL/Containers & Trailers/All Customers. Retrieved from Canadian National Railway Company Web site: http://ecprod.cn.ca/ebusiness/eDistribution/english/public/PriceDocumentDisplay
Canadian National Railway Company. (2012). Glossary of Terms. Retrieved from Canadian National Railway Company Web site: http://www.cn.ca/en/corporate-citizenship-safety-glossary.htm
Canadian Pacific. (2012). Facility List. Retrieved from Canadian Pacific Web site: http://www.cpr.ca/en/our-network-and-facilities/Pages/FacilityList.aspx?FacilityType=Intermodal&ProvinceState=British Columbia (BC)
Cap-Marine Assurances & Réassurances SAS. (2011). Shipping and Shipbuilding Markets 2011 Annual Review. Retrieved from Barry Rogliano Salles Web site: http://www.brs-paris.com/annual/annual_histo/annual_review_2011-a.pdf
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
44
Cariou, P. (2011). Is slow steaming a sustainable means of reducing CO2 emissions from container shipping? Transportation Research Part D: Transport and Environment, 260-264.
CSX Transportation, Inc. (2012). Carbon Calculator v2. Retrieved from CSX Transportation, Inc. Web site: http://www.csx.com/index.cfm/customers/tools/carbon-calculator-v2/
CSX Transportation, Inc. (2012, August 17). Customers. Retrieved from CSX Intermodal Web site: http://www.csxi.com/share/csxicustomer/main/docs/International_Service_Matrix-REF21399-REF24494.xls
CSX Transportation, Inc. (2012). Intermodal Schedules. Retrieved from CSX Transportation, Inc. Web site: http://shipcsx.com/public/ec.shipcsxpublic/Main?module=public.ischedule
Dizikes, P. (2010, November 8). The 6-percent solution. Retrieved from MIT news Web site: http://web.mit.edu/newsoffice/2010/corporate-greenhouse-gas-1108.html
Environmental Defense Fund. (2012, February 17). Smart Moves: Creative Supply Chain Strategies Are Cutting Costs and Emissions. Retrieved from Environmental Defense Fund Web site: http://business.edf.org/sites/business.edf.org/files/smartmoves_07_screen.pdf
Florida East Coast Railway. (2011). About. Retrieved from Florida East Coast Railway Web site: http://fecrwy.com/about
Herbert Engineering Corporation. (2011, June). Carbon Footprint Study for the Asia to North America Intermodal Trade. Retrieved from Port of Seattle Web site: http://www.portseattle.org/cargo/green-gateway/documents/carbon_footprint_study_20110610.pdf
ICF Consultung. (2005, April). Assessing the Effects of Freight Movement on Air Quality at the National and Regional Level. Retrieved from US Federal Highway Administration Web site: http://www.fhwa.dot.gov/environment/air_quality/publications/effects_of_freight_movement/
Maersk Line. (2010, August 23). Slow steaming is here to stay. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=news&path=/news/story_page/10/slow_steaming
Maersk Line. (2012, March 5). Maersk Line Brings Flagship Service Powered by BNSF Railway to North America. Retrieved from Maersk Line North America Web site: http://www.maersk-nam-marketing.com/advisories/2012/03/maersk-line-brings-flagship-service-powered-by-bnsf-railway-to-north-america/
Maersk Line. (n.d.). About Us. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/about_us
Maersk Line. (n.d.). Containers. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/our_services/containers
Maersk Line. (n.d.). Rates. Retrieved from Maersk Line Web site: https://www.maerskline.com/appmanager/maerskline/public?_nfpb=true&_nfls=false&_pageLabel=page_rate_search
Maersk Line. (n.d.). Reducing greenhouse gas emissions. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/about_us/environment/reducing_gas_emissions
Maersk Line. (n.d.). Schedules. Retrieved from Maersk Line Web site: http://www.maerskline.com/appmanager/maerskline/public?_nfpb=true&_nfls=false&_pageLabel=page_schedules_location
Maersk Line. (n.d.). Transpacific. Retrieved from Maersk Line Web site: http://www.maerskline.com/link/?page=brochure&path=/routemaps/newnetwork/transpacific
Muller, G. (1999). Intermodal Freight Transportation. Eno Transportation Foundation, Inc.
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
45
Net Resources International. (n.d.). Alameda Corridor Freight Line, United States of America. Retrieved from railway-technology.com Web site: http://www.railway-technology.com/projects/alameda/
Nike, Inc. (2010). Corporate Responsibility Report FY07-FY09. Retrieved from Nike, Inc. Web site: http://www.nikebiz.com/crreport/pdf/
Norfolk Southern Corp. (2012, March 1). International Service Matrix. Retrieved from Norfolk Southern Intermodal Web site: http://www.nscorp.com/nscintermodal/Intermodal/Business/Business_Groups/International/Service_Matrix/international.xls
Oxford University Press. (2012). Dictionary. Retrieved from Oxford Dictionaries Web site: http://oxforddictionaries.com/?region=us
Pacer International, Inc. (2004, April 27). Investor Relations. Retrieved from Pacer International, Inc. Web site: http://phx.corporate-ir.net/phoenix.zhtml?c=107717&p=irol-newsArticle&ID=568736&highlight=
Petromedia Ltd. (2012). PortWorld Distance - Ship Voyage Distance Calculator. Retrieved from PortWorld Web site: http://www.portworld.com/map/
Thomas Publishing Company. (2012). Glossary of Supply Chain Terms. Retrieved from Inbound Logistics Web site: http://www.inboundlogistics.com/cms/logistics-glossary/
Union Pacific Railroad Company. (2012, September 17). Intermodal. Retrieved from Union Pacific Railroad Company Web site: http://www.uprr.com/customers/intermodal/attachments/intl_matrix.pdf
Union Pacific Railroad Company. (n.d.). International Intermodal. Retrieved March 2012, from Union Pacific Railroad Company Web site: https://www.uprr.com/customers/intermodal/intl/index.shtml
US Department of Commerce. (2009). Distances Between United States Ports. Retrieved from National Oceanic and Atmospheric Administration Web site: www.nauticalcharts.noaa.gov/nsd/distances-ports/distances.pdf
US Department of Transportation. (2011, January). America’s Container Ports: Linking Markets at Home and Abroad. Retrieved from Research and Innovative Technology Administration Web site: http://www.bts.gov/publications/americas_container_ports/2011/
US Environmental Protection Agency. (2008, May). Climate Leaders Greenhouse Gas Inventory Protocol Core Model Guidance: Optional Emissions from Commuting, Business Travel and Product Transport. Retrieved from US Environmental Protection Agency Web site: http://www.epa.gov/climateleadership/documents/resources/commute_travel_product.pdf
US Environmental Protection Agency. (2008, May). Climate Leaders Greenhouse Gas Inventory Protocol Core Module Guidance: Direct Emissions from Mobile Combustion Sources. Retrieved from US Environmental Protection Agency Web site: http://www.epa.gov/climateleadership/documents/resources/mobilesource_guidance.pdf
US Environmental Protection Agency. (2012, November 5). SmartWay Transport Partnership Freight Shippers. Retrieved from US Environmental Protection Agency Web site: http://www.epa.gov/smartway/documents/partnership/trucks/public-bin-export.xlsx
Wikipedia. (2009, February 23). File:BNSF Railway system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:BNSF_Railway_system_map.svg&page=1
Wikipedia. (2009, June 7). File:Canadian National System Map.PNG. Retrieved from Wikipedia Web site: http://en.wikipedia.org/wiki/File:Canadian_National_System_Map.PNG
Wikipedia. (2009, June 9). File:Canadian Pacific System Railmap.PNG. Retrieved from Wikipedia Web site: http://en.wikipedia.org/wiki/File:Canadian_Pacific_System_Railmap.PNG
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
46
Wikipedia. (2009, February 23). File:CSX Transportation system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:CSX_Transportation_system_map.svg&page=1
Wikipedia. (2009, February 23). File:Norfolk Southern Railway system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:Norfolk_Southern_Railway_system_map.svg&page=1
Wikipedia. (2009, March 18). File:Union Pacific Railroad system map.svg. Retrieved from Wikipedia Web site: http://en.wikipedia.org/w/index.php?title=File:Union_Pacific_Railroad_system_map.svg&page=1
World Economic Forum. (2009). Supply Chain Decarbonization: The Role of Logistics and Transportation in Reducing Supply Chain Carbon Emissions. Retrieved from World Economic Forum Web site: https://members.weforum.org/pdf/ip/SupplyChainDecarbonization.pdf
World Resources Institute. (2009, June). GHG Protocol tool for mobile combustion. Version 2.0. Retrieved from Greenhouse Gas Protocol Web site: http://www.ghgprotocol.org/files/ghgp/tools/WRI_Transport_Tool.xls
World Resources Institute. (2012, August). Emission Factors from Cross-Sector Tools. Version 1.3. Retrieved from Greenhouse Gas Protocol Web site: http://www.ghgprotocol.org/files/ghgp/tools/Emission-Factors-from-Cross-Sector-Tools-(August-2012).xlsx
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
47
Attachments
Volume and Mass
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
48
Distance
Emission Factors
Cost: 20-Foot Standard Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
49
Cost: 40-Foot Standard Container
Cost: 40-Foot High Cube Container
Cost: 45-Foot High Cube Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
50
Time
Emissions: 20-Foot Standard Container
Emissions: 40-Foot Standard Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
51
Emissions: 40-Foot High Cube Container
Emissions: 45-Foot High Cube Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
52
Multiple Objective Analysis Results: 20-Foot Standard Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
53
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
54
Multiple Objective Analysis Results: 40-Foot Standard Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
55
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
56
Multiple Objective Analysis Results: 40-Foot High Cube Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
57
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
58
Multiple Objective Analysis Results: 45-Foot High Cube Container
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
59
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
60
Sensitivity Analysis: 20-Foot Standard Container
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Emissions
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Time
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Ove
rall
uti
lity
Rail point-to-point price (US$/mi)
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
61
Sensitivity Analysis: 40-Foot Standard Container
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Emissions
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Time
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Ove
rall
uti
lity
Rail point-to-point price (US$/mi)
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
62
Sensitivity Analysis: 40-Foot High Cube Container
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Emissions
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Time
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
(0.2)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Ove
rall
uti
lity
Rail point-to-point price (US$/mi)
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
Case Study: Multiple Objective Analysis of Intermodal Freight Transportation Routes for REI’s Inbound Logistics
63
Sensitivity Analysis: 45-Foot High Cube Container
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Emissions
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.0 0.2 0.4 0.6 0.8 1.0
Ove
rall
uti
lity
Weight on Time
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15
(0.5)
0.0
0.5
1.0
1.5
2.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Ove
rall
uti
lity
Rail point-to-point price (US$/mi)
Route 1
Route 2
Route 3
Route 4
Route 5
Route 6
Route 7
Route 8
Route 9
Route 10
Route 11
Route 12
Route 13
Route 14
Route 15