Supply Chain planning at Philips Lighting Lumileds

Supply chain planning at Philips Lighting Lumileds

How certain do we like to be?

A design and implementation of a stock control model to balance customer service and stock levels in an

end to end environment to improve product availability.

Author:

R. Hartevelt

Public

Master thesis project – Roy Hartevelt Page 2

How secure do we like to be?


end to end environment to improve product availability.

Master thesis project

Bergen op Zoom, April 14 2011

Name Roy Hartevelt

University Delft University of Technology

Faculty Technology, Policy and Management

Program Infrastructure systems & services

Section Transport and logistics

Company Philips

Supervisory committee:

Prof.dr. L. Tavasszy TU Delft

Drs. J.H.R. van Duin TU Delft

Drs. H.G. van der Voort TU Delft

Ir. J-E Talsma Philips

Ir. H. Rulkens Philips


Management summary Improving service level at the lowest possible costs is and will always be one of the key objectives of

Philips Electronics. This research illustrates how a part of the Philips supply chain control is setup

/designed to support this objective. Within Philips, the business unit called Lumileds supplies LED’s to its

customers. One of the components used to manufacture LED’s is made at the component manufacturer

that is subject to this thesis.

The objective of improving service levels at the lowest possible cost can be enabled by supply chain

control. Of course superior service levels can be realized with excessive inventory levels. However that

ignores the objective of lowest costs, because inventory cost money. Therefore an optimal balance

between service level and inventory must be achieved. This balance depends on a number of different

drivers like lead-time, lead-time variability, manufacturing quality and demand pattern. This research

will define a model that generates advice to achieve the desired balance, taking into account all relevant

drivers.

Based on literature research and analysis of the current Supply chain performance between the

component supplier and the manufacturer an advice model is designed. This model generates specific

stock level advices but also generates insights in the Supply chain uncertainties and lead-times. Besides

the stock optimization part of the model, the model functions as a decision supporting tool for Supply

chain balancing and stock level considerations. To gain the maximum output of the model, intensive

cooperation between the component supplier and the manufacturer is a must.

The performed research answers on the main research question:

WHICH SUPPLY CHAIN PLANNING CONTROL IS NEEDED FOR A MOST SUITABLE STOCK SITUATION TO SECURE

THE SAFETY STOCK LEVELS BETWEEN A COMPONENT SUPPLIER AND LED ASSEMBLY & TEST

MANUFACTURER.

The advice models gives a product specific advice based on the different lead-times and the Supply chain

uncertainties including a specific demand pattern. The component supplier’s manufacturing unit

consists of a front-end and a back-end part of the line. The model calculates for the front-end of the

component supplier an advice based on a Kanban replenishment strategy. The back-end of the

component supplier is controlled via replenishment and managed by a Re-Order-Point (ROP) calculation.

The ROP and Kanban boundaries are set monthly, the replenishment orders are calculated weekly. As a

result of above structure the organization is able to measure the different parts of the Supply chain and

the overall Supply chain as well. The measurements about the sub parts of the Supply chain are a result

of the new control model. The models results in transparency and the opportunity to simulate the

impact of variance or lead-time reduction on stock levels. Based on the simulation results the model

demonstrates the improvement potential and proves that an equal service is possible with lower stock

levels. The influence of the component supplier and the LED assembly & test manufacturer can be used

to improve and focus on cycle time reduction and controlling the uncertainties in the Supply chain as


much as possible. Finally, the models are the first step for further improvements between the

component supplier and the manufacturer.

The component supplier and the manufacturer both have a drive to improve their Supply chain controls

and aspire an increased level of cooperation. This is in line with literature; suggesting that Supply chain

transparency, information sharing and an intensive relationship has positive influences on Supply chain

performance. This intensive relationship is decisive for this research but also for future improvement of

Supply chain processes.

During the pilot phase (1 month) the models were used with 2 product families. Based on the

experiences of the field experts modifications were made before we extend the pilot with another

month. During the second pilot 4 extra product families were included. Based on the results out of the

second pilot the models were finalized and handover to operations ready. The last part of the hand-over

to operations was to imbed the models in the standard organization. Term-Of-References (TOR) was

made for creating meeting guidelines and meeting inputs & outputs /results. With the help of the TOR a

structure was set for having efficient and fruitful weekly and monthly alignment meetings. As a direct

effect of the transparent meeting structure thinking in improvement opportunities were embed in the

day to day business easily.

The need to focus more on integral supply chain planning and optimization instead of local optimization

plans has ascended the past few years. There are some appropriate ways to determine an optimal

solution for the optimization challenge. Research has indicated required actions to optimize the LED

supply chain. Without actions there will be too high inventory levels and yield losses with resulting

financial losses. All these different kinds of behavior provoke business losses and will harm continuity. In

case industrial consumers decide not to purchase, major sales loss is incurred.

The advice models (Re-Order-Point, Replenishment and Kanban) are based on specific risks (lead-time

and uncertainties). The models generate advice based on these risks for a desired service level. The

desired service level is expressed in the Z factor. The models give product specific advice for safety

stock, economical stock levels and quantities to produce based on the requested service level. The

outcomes of the model have an average safety stock of 1.85 week at the assembly and test

manufacturer for a service level of 95%. For a service level of 99% a safety stock of 2.6 weeks is required

with a total average stock of 3.1 weeks. The results of the advice model (Figure 1) indicate that a

product safety stock level can optimize to the most suitable stock positions for the LED supply chain at

Philips: meaning high customer service with the right underpinning stock levels in the chain (at the right

balanced cost levels). The benefits of the Supply chain control model are enormous. Without the model

the Lumileds business (at 12 Million pieces sales per week) needs 9 times more stock (see Figure 1).


Figure 1 Results Supply Chain control model

Figure 1 shows the model calculation for the most optimal situation and the situation as it was without

the new control models. It shows the (safety) stock needed to cover the demand during lead-time

(vertical) and the sales per week in Million(s) (horizontal). Due to project effort the lead time of the

component supplier is optimized from 12 to 1.31 weeks and secondly, the supply variance is reduced

from 0.61 to 0.21 (in CV value).

The implemented supply chain control model has initially led to a 720 K Euro stock cost (Safety stock:

420 K Euro, Demand during Lead-time: 300 K Euro) reduction (based on 12 Million pieces sales per

week) with a design and implementation cost of 150K Euro.

-

5,000

10,000

15,000

20,000

25,000

30,000

35,000

1 2 4 6 8 10 12

Nm

br

pie

ces

(* 1

00

0)

Sales per week (M pieces)

Stocklevel effects control model

Dem LT. Old

Safety stock Old

Dem LT. New

Safety stock New


Preface This Master’s thesis is the final project for receiving the Master of Science degree in System Engineering,

Policy Analysis and Management with the specialization Logistics at Delft University of Technology. It has

been carried out at Philips between March and August 2010.

I specially want to thank my supervisors at Philips, Hubert Rulkens and Jan-Edzard Talsma, for their

advice, support and encouragement.

I would also like to thank my supervisors at Delft University of Technology, Ron van Duin and Haiko van

der Voort for their valuable advice.

Finally I would like to direct my warmest thanks to my wife (Annette) and our three little ladies’ (Elmyra,

Elena and Evely) for always supporting me.

Thank you!

Bergen op Zoom April 14, 2011

Roy Hartevelt


Table of contents

MANAGEMENT SUMMARY .................................................................................................................................... 3

PREFACE ................................................................................................................................................................ 6

TABLE OF CONTENTS .............................................................................................................................................. 7

INDEX OF TABLES AND FIGURES ............................................................................................................................. 9

FIGURES ....................................................................................................................................................................... 9

TABLES ....................................................................................................................................................................... 10

1.0 INTRODUCTION .............................................................................................................................................. 11

1.1 BUSINESS CHALLENGE .............................................................................................................................................. 11

1.2 RESEARCH PROJECT ................................................................................................................................................. 11

2 ANALYSIS OF SUPPLY CHAIN CONTROL ............................................................................................................. 19

2.1 INTRODUCTION TO SUPPLY CHAIN CONTROL ................................................................................................................. 19

2.2 WHAT IS SUPPLY CHAIN CONTROL? ............................................................................................................................ 19

2.3 EFFECTS OF SUPPLY CHAIN CONTROL ON THE MANUFACTURER AND THE COMPONENT SUPPLIER .............................................. 21

2.4 ROOT CAUSES AND INFLUENCES ON SUPPLY CHAIN CONTROL ........................................................................................... 21

2.5 SUPPLY CHAIN CONTROL MODELS ............................................................................................................................... 23

2.6 CONCLUSION SUPPLY CHAIN CONTROL ANALYSIS ........................................................................................................... 25

3 ANALYSIS OF THE COMPONENT SUPPLIER AND THE LED ASSEMBLY MANUFACTURER ..................................... 26

3.1 SUPPLY CHAIN MANAGEMENT ................................................................................................................................... 26

3.2 LED ASSEMBLY AND TEST MANUFACTURER .................................................................................................................. 29

3.3 STAKEHOLDER ANALYSIS ........................................................................................................................................... 29

3.4 CONCLUSION ANALYSIS ............................................................................................................................................ 36

4 ANALYSIS OF COMPONENT THROUGH-PUT-TIME ............................................................................................. 39

4.1 INTRODUCTION ...................................................................................................................................................... 39

4.2 FRONT-END ........................................................................................................................................................... 42

4.3 BACK-END ............................................................................................................................................................. 43

4.4 SUPERMARKET STOCK LEVEL CALCULATION ................................................................................................................... 44

5 DESIGN SETUP ................................................................................................................................................... 47

5.1 INTRODUCTION ...................................................................................................................................................... 47

5.2 PERFORMANCE MEASUREMENTS ............................................................................................................................... 47

5.3 UNCERTAINTY ........................................................................................................................................................ 47

5.4 DATA ................................................................................................................................................................... 51

5.5 CONCLUSIONS ........................................................................................................................................................ 52

6 MODEL DESIGN ................................................................................................................................................. 53

6.1 INTRODUCTION MODEL DESIGN ................................................................................................................................. 53

6.2 EVALUATION DESIGN OBJECTIVE................................................................................................................................. 53

6.3 MODEL DESIGN BASED ON THE THEORY ....................................................................................................................... 53


6.4 PRACTICAL DESIGN AND MODEL RESULT ...................................................................................................................... 53

6.5 RESULTS ............................................................................................................................................................... 58

6.6 CONCLUSION MODEL DESIGN .................................................................................................................................... 59

7 SUMMARY DESIGN AND MODELING PHASE ...................................................................................................... 61

8 FROM MODEL TO PRACTICE .............................................................................................................................. 64

8.1 INTRODUCTION ...................................................................................................................................................... 64

8.2 GENERAL USABILITY OF THE MODEL ............................................................................................................................ 64

8.3 USING THE MODEL FOR A PILOT ................................................................................................................................. 65

8.4 SUMMARY FROM MODEL TO PRACTICE ........................................................................................................................ 67

9 CONCLUSION AND RECOMMENDATIONS .......................................................................................................... 68

9.1 INTRODUCTION ...................................................................................................................................................... 68

9.2 CONCLUSIONS OF THE RESEARCH ............................................................................................................................... 68

9.3 RECOMMENDATION FOR FURTHER RESEARCH ............................................................................................................... 72

9.4 REFLECTION ........................................................................................................................................................... 73

REFERENCES ......................................................................................................................................................... 74

APPENDICES......................................................................................................................................................... 77

APPENDIX A PRODUCT FAMILIES ..................................................................................................................................... 78

APPENDIX B RELATION BETWEEN STOCK LEVEL AND SERVICE LEVEL ......................................................................................... 79

APPENDIX C PHILIPS ANALYSIS ........................................................................................................................................ 80

APPENDIX D DETAILS SUB PROCESSES FRONT-END COMPONENT SUPPLIER ............................................................................... 82

APPENDIX E DETAILS SUB PROCESSES FRONT-END COMPONENT SUPPLIER ................................................................................ 83

APPENDIX F MODEL RESULTS KANBAN FRONT-END COMPONENT SUPPLIER .............................................................................. 84

APPENDIX G RACI MODEL SUPPLY CHAIN CONTROL MODEL .................................................................................................. 85

APPENDIX H IDEF0 LEVEL 2 ........................................................................................................................................... 86

APPENDIX I MONTHLY RE-ORDER-POINT PROCESS ............................................................................................................. 89

APPENDIX J WEEKLY REPLENISHMENT PROCESS .................................................................................................................. 95

APPENDIX K TUTORIAL SUPPLY CHAIN CONTROL MODELS ................................................................................................... 102

APPENDIX L GLOSSARY OF TERMS .................................................................................................................................. 130


Index of tables and Figures

Figures Figure 1 Results Supply Chain control model ............................................................................................... 5

Figure 2 Potential LED market in (billion) US$ (as per 10-2010)................................................................. 11

Figure 3 research scope of the project ....................................................................................................... 13

Figure 4 design methodology (adapted from Herder and Stikkelmans, 2004)........................................... 16

Figure 5 Uncertainty in the supply chain (T. Davis, 1993) .......................................................................... 20

Figure 6 Inventory level with safety stock (S. Lutz e.a., 2003) .................................................................... 20

Figure 7 IDEF0 Supply Chain control model as-is situation ......................................................................... 23

Figure 8 Safety-stock calculation comparison ............................................................................................ 24

Figure 9 value chain LED production .......................................................................................................... 26

Figure 10 supply chain execution run book ................................................................................................ 27

Figure 11 performance current supply chain control model ...................................................................... 28

Figure 12 CV throughput time component supplier ................................................................................... 29

Figure 13 the LED worldwide supply chain organization ............................................................................ 30

Figure 14 the worldwide supply chain organization in functional areas .................................................... 31

Figure 15 stakeholders in the value chain .................................................................................................. 32

Figure 16 stakeholder diagram ................................................................................................................... 33

Figure 17 Simplified Supply chain of Lumileds ............................................................................................ 39

Figure 18 Distribution indentification throughput-time front-end ............................................................ 40

Figure 19 Front-end component supplier ................................................................................................... 40

Figure 20 Back-end component supplier .................................................................................................... 41

Figure 21 Throughput-time back-end component supplier ....................................................................... 41

Figure 22 Entry process steps assembly & test manufacturer ................................................................... 42

Figure 23 Median throughput time front-end ............................................................................................ 42

Figure 24 Median throughput time back-end ............................................................................................. 43

Figure 25 Pareto negative effects on throughput-time component supplier ............................................ 44

Figure 26 Z value development ................................................................................................................... 45

Figure 27 Causal model of uncertainty and customer performance .......................................................... 47

Figure 28 Concept advice model ................................................................................................................. 50

Figure 29 Planning of replenishment orders .............................................................................................. 51

Figure 30 Model design (IDEF0 level 0) ....................................................................................................... 54

Figure 31 Model design (IDEF0 level 1) ....................................................................................................... 55

Figure 32 ROP calculation model ................................................................................................................ 56

Figure 33 Weekly replenishment calculation ............................................................................................. 57

Figure 34 Planning rules and deploy control model ................................................................................... 59

Figure 35 ROP development financially ..................................................................................................... 60

Figure 36 Concept advice model ................................................................................................................. 69

Figure 37 Planning of replenishment orders .............................................................................................. 70


Figure 38 stakeholder diagram ................................................................................................................... 71

Tables Table 1 requirements, criteria, constraints & design option stakeholders ................................................ 36

Table 3 Average stock level advice for data set .......................................................................................... 46

Table 2 Replenishment strategies (P. Suwanruji ao, 2005) ........................................................................ 49

Table 4 Results correlation analysis stock level and CLIP ........................................................................... 79


1.0 Introduction

1.1 Business challenge LED (light emitting diode) is a fast growing product in many consumer and professional applications.

Main reasons for that can be found in the green image of the product, extended use in several

electronic devices and as replacement for current general lighting products. In more detail: the use of

LED is increasing in many (handheld) devices like the mobile phones, displays, remote controllers and TV

screens and brings a boost in industrial need of the new light source. Besides the ‘new’ market for the

LED product, the LED is the replacement for light sources in automotive lighting, aviation lighting and

traffic signals. All of these new demands will materialize in the coming decade.

Financially: the potential market for LEDs has expanded dramatically, from US$7 billion in 2009 to

US$10.7 billion in 2010, which is a growth rate unreachable by any other electronic product. Along with

growing LED brightness and falling prices, the share of LED in general lighting field is expected to be

increased greatly; the general lighting market is of huge potential with the market size reaching US$100

billion. Promisingly, the LED market can reach US$20.4 billion in 20121.

Figure 2 Potential LED market in (billion) US$ (as per 10-2010)

1.2 Research project

1.2.1 Research objective

The objective of this research is shaped by: internal and external business drivers, the business challenge

as described in par 1.1 and a business improvement program within Philips Lighting Lumileds.

Internal drivers are mainly cost driven. Focus areas are yield and utilization improvements, inventory

optimization and cycle time reduction. External drivers have a totally different scope and are shaped by

market circumstances. Market growth caused a ramp up in the industry. Secondly, the technology in the

1 Financial figures according the Global and China LED Industry Report, 2009-2010; ResearchInChina

0

10

20

30

2009 2010 2011 2012

in billion US$

year

Potential LED market


area of color control is moving fast. Product releases due to color control improvements are done at a

frequent and regular base. Finally, introducing the LED in the automotive industry brought besides a

market increase, also new quality standards.

The business challenge is to accelerate the internal and external drivers faster in comparison with

competitors. Philips Lumileds is one of the top three players in the high class technology market in Ultra-

High Brightness LED. For saving the future market position a strong focus is at high quality standards and

cost savings projects as part of the daily operations and the business strategy. As a result of the business

focus the BU management team started up an improvement program. One of the assignments within

the improvement program was to execute an assessment about an introduction of a new product

family. As an outcome of the assessment three main issues were detected in the areas of:

1. Components manufacturing,

2. Logistics control,

3. LED assembly and test.

This research will focus at improvement area 2 – Logistics control at Philips Lumileds. This challenge can

be translated in a problem statement:

Problem statement:

In order to serve the global LED market in a most suitable way Lumileds must continuously improve

their business in a structured way. Lumileds has a clear focus how to deal with product and market

uncertainties and translated this focus in a business wide improvement program. Insights are needed

to balance service- with stock levels and to find an optimum in the utilization (shop floor scheduling)

of the LED assembly and test.

Common to all manufacturing companies, regardless of size, type of product or manufacturing process is

the need to control the flow of materials from suppliers, through manufacturing and distribution to the

customer (G.C. Stevens, 2007). Also in our business configuration the role of supply chain is crucial. This

research will identify which criteria are relevant to determine economical stock values and deliver a

model for stock level calculation for the midterm (monthly) and the short-term (weekly). Those advice

models generate stock advice on the one hand based on customer demand and it’s variance and on the

other hand the supply and it’s variance. For the midterm, those models can help to determine further

tactical decisions. The result of the models (midterm and short-term) can help to take decisions for

these challenging considerations. This research investigates some of these future considerations and

there will be provided recommendations how to use the advice in practice and if the chosen calculation

method will fit in an optimal way for each phase in the product-life-cycle.

The research focus is clearly on improving service level and optimizing stock levels. The model must be

used and understood by several business lines of Philips Lumileds. These business lines will have

different backgrounds, goals and work levels. Attention to this multi actor setting is important and

frequent contact and information sharing will contribute to this. Recognizing the multi actor setting for


stock level adjustments is important. The components supplier and the LED assembly and test

manufacturer both have the intention to optimize stock levels in a structured and mutually agreed way.

The goal of the assembly and test manufacturer is creating a demand/supply advice for the component

manufacturer and gets insights on the supply chain risks from the models, at the other hand the

component supplier owns and controls their own production process and internal optimization. For

model validation the model will be tested by experts in the field from both actors.

Research objective:

Generate, implement and deploy supply chain planning controls for shop floor scheduling and material

replenishment for the LED products between a component supplier and LED assembly and test

manufacturer.

The problem statement and research objective make explicitly that the focus of the research is on stock

level optimization, utilization in relation with service levels.

1.2.2 Research boundaries

Philips practical influence stops at the customer’s factory or DC. The focus of the research is represented

in Figure 3 research scope of the project. Important to mention is that this research will not covers the

integral supply chain of Lumileds. The research areas are the relations and independencies between LED

assembly and test and a key component manufacturer with a focus on the internal drivers as mentioned

in par 1.2.1.

The external factors are regarded as stable for the short term. In the long term the boundaries can easily

change due to the LED market situation as we have today. This research focuses on the sales figures of

the assembly and test (promotions included). The temporary shifts in the demand with the disturbances

for the total chain should be covered in the solution design. When stocks and lead times are reduced but

losses do not outweigh the gains the advice is not successful. A balance must be found between stock

level investment and service level (Jammernegg and Reiner, 2007).

Figure 3 research scope of the project

LED assembly and test

Component manufacturer


Information sharing can significantly improve the performance of the supply chain (Zhao, 2002).

Transparency and information sharing is necessary to gain a complete vision about the risk in the whole

supply chain and fit different parts into one consistent chain. Figure 3 shows the boundaries of the

research but also explicitly mentions the crucial transparency. The model not only gives the component

manufacturer production advice but also a convincing and informative role. Potential users must trust,

understand and be able to use the model, before the advice can be considered successful. Firstly the

interactions between the involved actors will be examined, to create trust and understanding. Frequent

contact with and clarification for the stakeholders is probably an outcome to handle this multi actor

complexity. Beside these frequently updates workshops can be a useful tool to create the required

comprehension and trust.

1.2.3 Research questions

To achieve the research objective stated in section 1.2.1 the following research questions are

formulated:

Research question

WHICH SUPPLY CHAIN PLANNING CONTROL IS NEEDED FOR AN MOST SUITABLE STOCK SITUATION TO SECURE THE

SAFETY STOCK LEVELS BETWEEN A COMPONENT SUPPLIER AND LED ASSEMBLY & TEST MANUFACTURER.

In order to find an answer on the main research question several sub-questions are formulated:

1. Which facts determine the most suitable supply chain planning method?

Method: Literature research of scientific publications (main search criteria: demand planning, supply

planning, Supply chain design, Supply chain planning strategies), internal Philips documents & field

research.

Result: Insights in the factors influencing the current supply chain planning method significantly. The

influence can be used as input to determine key selecting criteria for selecting alternative planning

methods.

2. Which planning model is most suitable for the LED supply chain?

Method: Literature research of scientific publications (main search criteria: Supply chain design, semi

conductor industry, planning strategy), internal Philips documents & field research.

Result: Insights in the factors influencing the supply chain planning. The influence can be used as input

to determine the most optimal supply chain planning model.

3. Which processes are related to the supply chain planning processes and what are the

interdependencies?

Method: Literature research of scientific publications (main search criteria: planning processes, supply

chain planning, supply and demand match) and field research.


Result: A transparent process description. Mutual understanding and agreement about the business

processes and the independencies between several sub processes is necessary to make the insight of

the model clear and accepted and acceptable to use.

4. What are the relevant actors, what role do they currently have and what role should they have

according the new model?

Method: Field research, interviews with stakeholders, Literature research of scientific publications (main

search criteria: process design, organization setup).

Result: Revaluation of the current used stock control models and advice to adapt or maintain this

procedure, based on the new insights on economical stock calculation for different product families.

5. Which key performance indicators are relevant to evaluate the supply chain plan model and

what are the critical success factors?


search criteria: performance indicators in the Supply chain, performance indicators design

/implementation).

Result: A list with the most relevant supply chain performance indicators and secondly a risk profile in

which all the important factors influencing the supply chain planning performance (be. Stock levels,

lead-time, variance) are mentioned. The risk profile is necessary to evaluate the calculation rules as used

in the model and find general conclusions about the behavior of the model within the different product

life cycle faces.

6. What actions are recommended to convince the actors about the benefits of the new supply

chain planning model and way of working?


search criteria: actor management, change management, soft controls).

Result: An actor analysis in which all actors are described with all relevant arguments (pro /con) about

the new planning model and way of working. The actor profiles are necessary to evaluate the difference

in change acceptance processes and are an input for the advice model.

1.2.4 Methodology of the research

This chapter defines the methodology that will be used to gather an answer on the formulated research

questions. The methodology is a framework for the research and can be used as a guideline in the

research activities. The design method used in this research is based on the complex multi actor and

multi requirements methodology (Herder and Stikkelman, 2004). The conceptual model is based on

IDEF0 diagrams. The IDEF0 diagrams are based on the outcome of several workshops together with the

field specialist. The output of the model construction will be visualized and presented to the


stakeholders. Based on the discussion during the presentation a pilot will be setup. The pilot will provide

insight information to optimize the model in the 2nd design.

Figure 4 design methodology (adapted from Herder and Stikkelmans, 2004).

One of the deliverables of the project is to deliver a tool and guidelines with the goal to optimize the

stock levels and to introduce supply chain controls. The decision support tool will enable management

to make the right decisions before we make the wrong one (R.E. Shannon, 1998).

The goals of the research are determined by the stakeholders (chapter 3). A clear formulation of goals is

essential to create suitable and accepted design and control afterwards if the design fits the desires of

the stakeholders. The goals are translated into objectives and constraints. These objectives and

constraints are matched with the design space. The design space is formed by the boundaries of the

specific situation. The research boundaries in combination with time and data restrictions determine the

design space for this research. With respect to the design space first a conceptual model is designed

(chapter 4). This conceptual model is discussed with experts and stakeholders and when necessary

adapted. After adapting the input of the field experts and stakeholders the conceptual model will be

translated in a real model (chapter 6). This model is tested and discussed again with stakeholders and

experts. After this evaluation the final model is designed and steps to use it in practice are provided

(chapter 8). Finally, results are used to test how good the model fits with the design objective.

1.2.5 Structure of the report

This report is split up in three phases: 1. Analysis, 2. Design and modeling and 3. Evaluation and

Validation. The three phases have an introduction, conclusion are split up in different chapters.


The analysis phase concerns the current state (chapter 2 & 3). In this phase the requirements and

decision space are formulated based on the analysis of the as-is situation, input of field-experts and

strategic knowledge about the direction as set by Lumileds senior management. This analysis will also

draw attention to the multi actor setting of the problem. The tensions between the actors will discussed

and options to handle these differences are formulated.

The second phase, the design and modeling phase, starts with an analysis of an optimal stock level

calculation method for a key component of the LED. This method is based on product life-cycle

characteristics which determine delivery performance (start with chapter 4). The profiles per product

family are input for the next level of detail as used in the short-term planning tool. The objectives and

constraints form the solution space and are together with the risk profile input for the design. Besides

this input a list of criteria will be formulated to evaluate the model. The solution space is an input for the

design of the model. The results of the model and the implications for the component manufacturer and

the LED assembly are described in chapter 5. With the input of the previous chapters the design phase is

descript in chapter 6. How to implement a stock optimization model is answered in chapter 7.

The last phase (phase 3) is the evaluation and validation phase and starts with a validation of the model

with a simulation (chapter 8). Conclusions and recommendations of this research are placed in chapter 9

including the validation of the research questions and other reflections about theory and practice.


Analysis Phase

The analysis phase is the first part of the design process. In this phase the concept and implications of

supply chain control and the structure of and relations between the component manufacturer and the

LED assembly & test are analyzed in detail. Absorbing uncertainties in the supply chain by well

positioned and calculated stock levels at indifferent positions in the supply chain are translated into

design parameters as input for the design and modeling phase. The results of the analysis are translated

in design requirements and objectives and are used for identification of the solution space.


2 Analysis of supply chain control

2.1 Introduction to supply chain control This chapter explains the importance of supply chain planning and control focused on a component

supplier and the LED assembly and test manufacturer. Besides the introduction this chapter is divided

into 5 stages, stage 1 (par. 2.2) introduction of controls in the supply chain and it indicates the

importance of this subject and the related problems and gives an overview of the most important and

interesting academic literature and conclusions. The second stage (par. 2.3) discusses the problems and

losses for manufacturers and component suppliers. As a result of stage 1 and 2 stage 3 brings a

discussion about root causes and effects of limited supply chain control. The fore last paragraph

introduces the different ways of supply chain control models. The last paragraph of this chapter starts a

discussion on the design phase for the setup of solution model to identify the optimal solution.

The analysis phase is the first step in the model description as provided in Figure 4. The analysis provides

a basic building block for understanding the business issue. Secondly, the insights gained in the analysis

supply an input for the design phase. The analysis follows the structure of the TIP approach

(Koppenjan&Groenewegen, 2005). TIP categorizes the system aspects into three sections.

T Technology; analyzes the content of the issue

I Institutions; analyzes the important stakeholder requirements

P Process; refers to the interplay of stakeholders and their interests

2.2 What is supply chain control? Doing a good job in supply chain starts with living with uncertainties. In Supply chain control means

coping with several uncertainties of all activities in the scope of business, like variance in customer

demand, supplier performance and manufacturing and last but not least it means to follow the demand

pattern of our customers (T. Davis, 1993). At the end of the day management is most interested in

customer satisfaction and low inventory levels.

Uncertainties in the supply chain are daily business; beating the uncertainties will never result in

satisfaction or in a different way an uphill battle will never end in a great victory. Living with uncertainty

in Supply chain means integrating variances of each part of the process into a stock calculation model.

Figure 5 shows the supply chain uncertainty as published by T. Davis in 1993. Unfortunately, not all

uncertainties can be eliminated. However, other initiatives can redesign the Supply chain to reduce the

impact of uncertainties (T. Davis, 1993).

How to live with uncertainty? This question can be split in a longterm and a shortterm answer. For the

longterm a supply chain modeling methodology can adress the network, interdependencies and other

more structure related issues. For the shortterm, we will use models with scheduling algorithms to fine-

tune day to day performance.


Figure 5 Uncertainty in the supply chain (T. Davis, 1993)

The real challenge is to manage the uncertainties in the supply chain in relation with customer service

level without incurring an undue burden of cost. Balancing the activities in the chain is the main goal to

achieve the necessary balance between cost and customer service.

Inventory has a stabilizing effect in supply chains (M.P. Baganha a.o., 1996) as a buffer to absorb

demand variability. The stabilizing function of stock will result in a lower variability in production than

the variability in market demand.

Figure 6 Inventory level with safety stock (S. Lutz e.a., 2003)

Figure 6 shows the development of inventory over time for a process with input and output quantities.

In case no uncertainty is involved the mean inventory can be lowered with the safety stock. With the

safety stock (SS) the described process can guarantee the deliveries. The mean inventory level is the

sum of the safety stock and half the quantity of input (lot size) into stock necessary to cover the usage

during lead-time (S. Lutz e.a., 2003).

The next step into supply chain control after understanding the supply chain uncertainties; translate all

known uncertainties into an inventory stocking policy. The constantly changes, so the uncertainties are

constantly changing too. Thus the inventory stocking policy is a dynamic process. Suppliers delivery

reliability can become better or worse with the time. On the other hand, demand for some products

becomes more predictable as products mature; demand for other products becomes more

unpredictable.


In creating a dynamic stock policy we commonly use generic settings for A/B/C stock-keeping units. The

classification of items by transaction volume does not necessarily reflect the combination of all

uncertainties of the total supply chain (H.L.L. Lee e.a., 1992). In the default setup of the model the

possibility should be there to differentiate the service levels based on a potential ABC classification.

Currently no classification is made due the fact that all products are rather new and product launches

are planned twice a year. For managing the long term and the day to day business, more complex

calculation models are necessary for creating the most optimal and suitable supply chain for the LED

business. Consequences of calculating stock levels and customer deliveries with wrong uncertainties

results in overstock on some items but under-stock on others. Miscalculating the lead times for material

movement along the supply chain could result in investing in the wrong resources for performance

improvement.

2.3 Effects of supply chain control on the manufacturer and the component

supplier The value stream network is relatively small compared with other networks in the high tech industries,

but nevertheless the limited level of alignment between the key component supplier and the assembly

and test manufacturer can be marked as critical. The complexity of coordinating the supply chain is

exploding for the assembly and test manufacturer due to the introduction of new products and

demanding customers in a booming market. The ever-shortening product life-cycles combined with long

lead-times are challenging for the planning department at the manufacturer besides the supply side

with their struggle to follow the demand of their customers. A certain portion of the total stock is

reserved for covering the demand during lead-time. Therefore decreasing lead-time is the right lever to

cut inventories. The second portion of the inventory is reserved for the safety stock; the safety stock is a

function of the service level, the demand uncertainty, the replenishment lead-time and the lead-time

uncertainty. For a ‘fixed’ service level there are three levers that affect the safety stock; demand

uncertainty, replenishment lead-time and lead-time uncertainty. This thesis focuses on modeling the

right calculation method between the key component supplier and the assembly and test manufacturer.

As a result of the fragmentation of the supply chain in independent entities that tends to optimize

locally instead of coordinating and optimizing the entire supply chain. Competitive advantage of a

multinational is lost or gained by how well a company manages a dynamic web of relationships that run

throughout its chain of suppliers, distributors and alliance partners (de Kok e.a., 2005).

2.4 Root causes and influences on supply chain control Managing uncertainty in the LED supply chain has currently a main focus on the lead-time and the lead-

time variance of the key component supplier. Current implemented models are calculating the demand

versus the supply. Based on calculations and experience, a production order is placed. No uncertainty is

taken into account in the current planning models. The current supply chain control based on demand is

controlled by the S&OP (Sales and Operations Planning) of the assembly and test manufacturer. In figure

7 the current supply chain control model is visualized with the IDEF0 (Integration DEFinition for

Function) method. This model illustrates that the control parameters are limited (capacity constraints


and carrier lead-time). This translates directly into limited possibilities to control the supply chain by

managing the uncertainty and stock levels. Between the planning setup of the key component supplier

and the assembly and test manufacturer there is a missing link between the short term and medium

term planning. The medium term planning cycle based on the sales and operations planning (S&OP) of

the assembly and test manufacturer is the starting point for determining the needed production

capacity. The short term planning cycle has a decentralized focus and is disconnected from the midterm

planning cycle. The key component supplier has independent weekly planning cycles calculating material

requirements based on component consumption. These disconnected processes cause long information

lead times and distortion. The long information lead time is caused by the several independent planning

cycles with a monthly update via the S&OP cycle. As an average half of the monthly period is a waste of

information lead time which can result in additional obsolescence risks. The focus of the logistics

manager of the assembly and test manufacturer is at reduction of replenishment lead-time from the key

component supplier and the variability of lead-time.

The IDEF0 (Structured and Design Technique, D.T. Ross, 1981) of the current supply chain control is

displayed in Figure 7. In step A1 the master production schedule (MPS) is setup based on the sales and

operations planning (S&OP) and the distribution of the components. Based on the MPS of the LED

assembly & test the component supplier will generate a material requirement plan (MRP) (A2). The

capacity constraints of the components production is the control element. The capacity constraints

control element takes into account the available production capacity and the factory supplies for the

long term (be. Investments for capacity increase). As input for the MRP the factory yield of the back-end

processes are taken into account. The work-in-process (WIP) is taken as a point of departure for the new

plan cycle.


Figure 7 IDEF0 Supply Chain control model as-is situation

The next step after the MPS and MRP is to align the factory plan (A3) with the demand of the assembly

and test manufacturer, a weekly schedule of a day to day plan. Al the steps after activity A4 are

transport and order-receiving at the LED assembly & test factory.

2.5 Supply chain control models Control of the supply chain means understanding and managing the uncertainty in the scope of the

chain. In a simplistic way the cycle stock (demand during lead-time) will be increased with a certain

percentage or some extra days of inventory. For calculating the safety stock levels as part of a

replenishment strategy model several variants are available to use, the question is which calculation

method of the safety stock is the most suitable for the LED situation and what are the differences and

explain them. Based on publications of S. Chopra eo, 2004, P. Suwanrauji eo, 2005 and S. Chandandeep

2010 three safety stock calculation methods have been selected:

SAFETY STOCK CALCULATION 1 (CHARLES ATKINSON, 2005);

[1]

SS1 = safety stock calculation according Charles Atkinson

Z = service level

LT = Average lead-time

δdem2 =Standard deviation of demand^2

TITLE:NODE: NO.: 0As is A0 SADT Lumileds

S&OP

A1

Prepare MPS

Png

A2

Generate MRP

Mhz

A3

Align Factory

planning back-

end Mhz

A6

Stock update

Penang

Platelet stock Png

A5

Prepare In-

transit overview

Yield back-end Mhz

WIP Mhz

Platelet distibution

MES

Capacity

constraints

Front-end Mhz

Carrier

Lead-time

Stock overview

Png

A4

Generate report

WIP back-end

MhzStock pre grinded wafers

Capacity

constraints

JD Edwards stock data

MPS –

Long term

MRP -

Short term

Confirmed

Planning

WIP

report

Goods

in-transit


dem2 = Average demand ^2

δLT2 = Standard Deviation of Lead Time ^2

SAFETY STOCK CALCULATION 2 (KENT LINFORT, 2006);

[2]

SS2 = Safety stock calculation according Kent Linfort

σFE2 = Forecast error

LTI = Lead-time interval

FI = Forecast interval

σLT2 = Lead-time error

D2 = Average demand during lead-time^2

Z = Desired service level

OCI = Order cycle interval

SAFETY STOCK CALCULATION 3 (DAVE PIASECKI, 2009) ;

[3]

SS3 = Safety stock calculation according Dave Piasecki

δ = standard deviation demand

Z = Service factor

LT = Lead-time factor

OC = Order cycle factor

Dem = Forecast-to-mean-demand factor

Figure 8 Safety-stock calculation comparison

With the help of a small example the three calculation methods (SS1, SS2 and SS3) are compared. The

result of the comparison is visualized in Figure 8, which shows a safety stock level at three different

service levels. The Lumileds Supply Chain needs Safety-stock for absorbing the uncertainties at the

demand and supply. In making the right choice related to the safety-stock calculation it is most

-

50

100

150

200

250

300

350

Z=1.64 Z=1.88 Z=2.33

SS in

Pcs

Safety-stock calculation comparison

SS1

SS2

SS3


important that the calculation method is able to deal with the different types of Supply Chain

uncertainties. The limited possibilities in SS3 to integrate the different uncertainty factors drop out this

calculation method. Calculation method SS2 needs a lot of not standard available information and

secondly the calculated safety-stock level is very close but higher than the SS1 calculation method.

Based on the characteristics (fast growing and many different uncertainties) of the LED business of today

the safety stock calculation of C. Atkinson (SS1) has the best fit and will be used in further stock level

calculations.

2.6 Conclusion supply chain control analysis The need to focus on supply chain control drastically increase due to the expanded product portfolio of

the LED products. The market is growing very fast and the competitors are aggressively competing for

market share. LED industrial customers require high service standards: Out of stock situations are not

tolerated. The direct impact of supply chain controls in out of stock improvement actions can be

significant. This project will deliver a model design and implementation and will open new ways to

improve or to minimize out of stock situation caused by supply chain controls (sub research question 1

and 2).

There are multiple root causes for out of stock situations in the supply chain of Lumileds. The most

important and relevant root causes for this research are: underestimation of uncertainty (production

lead-time and production quantity related), long order lead-time, demand underestimation, new

product introductions and data inaccuracy. All these root causes haves an effect on the supply chain

control model and will be incorporated in the design and modeling phase (sub research question 3)

Measuring stock levels can be done in several ways, the question is which one is the most suitable for

their business situation and how the calculation method can be integrated into the overall control

model. Finally a control mechanism is needed to control the supply chain between the component

supplier and the assembly and test manufacturer to serve the fast growing LED market in a for Philips

optimal way (main research question).


3 Analysis of the component supplier and the LED assembly

manufacturer The highlights of the Philips Company, within this research (Philips, Lighting and Lumileds) are

positioned in appendix C. The analysis on supply chain control and organization starts in paragraph 3.1.

In this chapter also the underlying reasons for this research are made explicit. After a brief introduction

and description the processes within and between the component supplier and the assembly and test

manufacturer are discussed (section 3.2). The multiple stakeholders all have other interests, goals and

influences in relation to supply chain planning optimization. Mapping this information (section 3.2) can

help to improve the process. Section 3.3 provides round up and summaries around these aspects.

Section 3.4 finally gives a wrap up of the whole analysis phase.

3.1 Supply chain management In a fast growing business environment supply chains are facing some typical pitfalls and opportunities

(H.L. Lee 1992). Pitfalls like not having the right metrics, a misunderstanding about what is customer

service and even more basic understanding who is the customer and finally limited and poor supply

chain coordination are the backbone of the analysis.

The value chain of LED is build up out of 3 blocks, see Figure 9. The first block “component production”

covers the component manufacturing in Maarheeze (NL) and Singapore (SI). Transport is executed by an

express courier. The LED assembly & test is located in Penang (MA). The LED assembly & test is driven by

sales orders. Activities done in Singapore are out scope of this research.

Figure 9 value chain LED production

3.2.1 Current supply chain processes

In Figure 10 shows the current run-book. Key elements in this run-book are the coupling between the

monthly update from sales and operations planning (S&OP) with the material requirement planning of

the LED assembly & test factory. The MRP is translated towards the component supplier. The back-end

factory planning is the fundament for next steps in the component factory.

The monthly S&OP update and the weekly alignment via the MRP with the component supplier have at

least a delay of one week in information retrieval and transmission. The monthly and weekly alignment

Component production

TransportLED

assembly & test


sessions are also discourage short production planning cycles leading to gross forecast errors and

inventory and backorder accumulation (L.L. Hau, 1992).

Besides sharing of information, coordinating of information is an important element in optimizing

Supply chains (W. Jammernegg, 2006). In the supply chain of LED the ownership of the component

stocks is at the LED assembly and test manufacturer. In running business the planner of the component

supplier is managing the stock levels and takes a decision if applicable (experienced based).

Figure 10 supply chain execution run book

As a result of shared responsibilities there are no suitable performance measures for the complete

supply chain. A reason for not having those kinds of metrics is that they are not directly linked to

customer satisfaction. As shown in Figure 11 the planned versus the real delivered quantities are

measured besides the confirmed line item performance (CLIP=yellow line) and the confirmed volume

performance (CVP=grey line). Not taken into account are the total order cycle time, average backorder

levels, average lateness or earliness and back order profile (backorders that are one week late, two

weeks later). The main focus at the component supplier was to produce the forecasted quantities during

the selected period. Most of the time the volume was there, a minor element is that the mix of the

products to be delivered was not in line with expectation. The plan performance based on volume (CVP

=grey line) was 90% during the selected period; the performance based on the mix (products delivered

versus ordered) was around 40% (CLIP=yellow line) during the selected period. Main reason for having a

focus on total quantity instead of delivering the right mix is difficult to answer. After several interviews

with field experts the overall conclusion was that the day target asked for a minimum number of moves

at a workstation. After beating the target of the required number of moves management was satisfied,


without taken any notice about the right moves. In this kind of situations the workforce will do were is

asking for; making moves!

Figure 11 performance current supply chain control model

Two focus areas in a supply chain area are the reduction of the replenishment lead-time from suppliers

and the variability of this lead-time (S. Chopra, 2004). In supply chains with large variability in lead-time,

reducing variance have a greater impact on reducing safety stock levels than cycle time reduction. In our

research we recognize that cycle time and the variance are both increasing. In phase 2 of this thesis we

focus on the three levers that affect the safety stock (demand uncertainty, replenishment lead-time and

lead-time variance).

Figure 12 shows de coefficient of variance (CV) of the throughput time of the component supplier per

period (CV calculation by S. Chopra et al, 2004). A period is defined as a week in 2010. The CV value is

increased by almost 100% in a period of 10 periods. The increase of the CV is caused by a growing

(standard deviation) δ and μ (mean) and on the other hand due to diffused production priorities

because of several new products in production. Secondly, balancing capacity between the component

supplier and the assembly and test manufacturer is done during the monthly Sales & Operations Plan

meeting. Deviation from plan due to higher sales volume as forecasted will directly effects the CV in a

negative way (increase).

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

0

500,000

1,000,000

1,500,000

2,000,000

2,500,000

3,000,000

3,500,000

4,000,000

4,500,000

5,000,000

wk1

00

1

wk1

00

2

wk1

00

3

wk1

00

4

wk1

00

5

wk1

00

6

wk1

00

7

wk1

00

8

wk1

00

9

wk1

01

0

wk1

01

1

wk1

01

2

wk1

01

3

wk1

01

4

wk1

01

5

wk1

01

6

wk1

01

7

wk1

01

8

wk1

01

9

wk1

02

0

wk1

02

1

wk1

02

2

wk1

02

3

wk1

02

4

wk1

02

5

wk1

02

6

"Co

nfi

rme

d L

ine

Ite

m P

erf

orm

ance

-C

LIP

& C

on

firm

ed

vo

lum

e

pe

rfo

rmac

e C

VP

(%

)

Pla

nn

ed

/ R

eq

uir

ed

vo

lum

e (

qty

)

Planned qty Achieved qty

CVP (%) actual CVP (%) target

CLIP (%) actual CLIP (%) target

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 2 4 6 8 10 12

Co

eff

icie

nt

of

Var

iati

on

Period

CV

Linear (CV)


Figure 12 CV throughput time component supplier

CV = δ/μ [4]

CV = coefficient of variance

δ = standard deviation

μ = mean

Managing the customer satisfaction with a better supply chain performance starts by managing the

variances at the upstream part of the chain. A part of the solution model in phase 2 will take care about

the upstream part of the chain – the component supplier.

3.2 LED assembly and test manufacturer

3.2.1 General introduction LED assembly and test manufacturer

The LED assembly and test manufacturer of Philips Lighting is located in Penang (Malaysia). The two

main suppliers of key elements are a Philips component supplier in Singapore and Maarheeze. The

components from Singapore are the most expensive ones and for that reason marked as leading in the

assembly factory. The leading position of Singapore brings Maarheeze into the position of a follower.

Key elements of a follower are that flexibility and production capacity should be no issue. However,

currently Maarheeze is not able to fill the pipeline with key components due to capacity issues. As a

result of this situation the assembly and test manufacturer is not able to follow the demand of our

customers.

This research will have a focus on the improvement possibilities between the component supplier and

the assembly and test manufacturer. The leading position of the components from Singapore is taken

into account in the analysis done in the next phase.

3.2.2 Ramp up in a fast growing market

Due to the recovery of the global economic market the LEDs is widespread use as the backlighting units

not only of large-sized LCD panels used in TV and computer screens but also on smaller LCDs in a broad

range of devices including notebooks, cell phones, portable navigation devices, keypads and many other

applications. For 2010 till 2015 the global LED market expects a double digits growth

(www.semiconductor-today.com), this double digit growth emphasized the need for controlling the

supply chain.

As a result of the fast growing LED demand a shortage is expected on the market by 2011. As a reaction

on the market shortage vertical integration at our competitors is going on for maximum safeguarding

their needed capacity.

3.3 Stakeholder analysis The stakeholder analysis has three main parts. Firstly, the supply chain organization (section 3.3.1.) of

the LED business unit and the position of the strategic plans. Secondly, the interests and objectives of

http://www.semiconductor-today.com/


the several stakeholders in the value chain (section 3.3.2/3). And finally, organizational acceptance and

the role of the project principal are discussed (section 3.3.4/5).

3.3.1 Supply chain organization

The world wide supply chain manager (WW SCM) is part of the management team of the business unit

(BU). In this position the WW SCM has a powerful business position to drive the BU into the direction of

the DIMEs project goals. The director customer service, director APAC (Asia Pacific) supply and the world

wide program manager are reporting to the WW SCM, see also Figure 13. SCM fulfills an important role

in the BU, in this setting the BU management prevent themselves from product and process design

without any supply chain consideration (H.L. Lee, 1992). This setup results that all of the anticipated

savings done by product design, manufacturing and assembly are not be lost by increased distribution

and inventory costs. Potentially costs increases are involved due to SCM intervention in the design,

manufacturing and assembly processes at the short term. At the long run the company benefits from an

optimal flexible supply chain concept for meeting demand. Flexibility is especially important in new

product introduction and ramp up, where demand is highly variable as well as unpredictable. The colors

in chart Figure 13 will be used in Figure 14, based on that the functional areas in the supply chain will be

discussed.

Figure 13 the LED worldwide supply chain organization

Every organizational layer in the three tier management organization of SCM has own tasks and

responsibilities to perform most optimal in design, manufacturing, assembly and sales processes.

Demand and supply processes are visualized in Figure 14. Customer service is responsible for the

customer interfaces, order handling, allocation processes during scarcity and customer specific demand

solutions, all demand related. Order scheduling activities at supply fulfillment close the loop between

customer demand and order fulfillment.

WW SCM

Director

Asia Pacific

Supply Chain

Planning

Manager

Director

Customer

Services

Logistics

Manager

Director

WW

Programs

Supply

Chain Exc.

Manager


Figure 14 the worldwide supply chain organization in functional areas

The APAC supply manager has overall responsibilities in demand and supply. The customer complexity is

translated via a Sales and operations planning (S&OP) process into a supply plan. Based on the S&OP (an

outlook of 1.5 year ahead) the midterm and short term planning activities can be aligned for meeting

the demand as efficient as possible. S&OP is also used for longer term investment decisions. The

logistics and planning managers are running the day to day business and balancing demand and supply

variance versus customer lead-times and stock management.

An important role in the APAC supply organization is the supply chain excellence department. This

department initiates and executes improvement projects in the total LED supply chain. Key supply chain

topics which correspond to the strategies for improvement in supply chain design, integral supply chain

performance measurements and integration of control and planning support systems are covered and

managed by the supply chain excellence manager.

3.3.2 Stakeholder setting

Figure 15 shows that in the component production many stakeholders are available. The many

stakeholders at the component production will compete for required capacity in achieving individual

goals. As mentioned in the introduction of this master thesis, the LED business is growing fast. In a fast

growing market development and engineering is an enabler for further business growth. Besides

manufacturing capacity, engineering and development activities require factory capacity. An optimal

supply chain planning model will bring stability and clearness for involved actors. Based on customer

forecasted demand all capacity left can be used by engineering /development.

DEMAND

PLANNING FULFILLMENT

Asia Pacific Supply Management

PlanningOrder Confirmation/scheduling

Key Customer Interface

Scenario Planning

Inventory Monitoring

Weekly Build Plans

MRP

Supply Chain ExcellenceS&OP Planning

Scenario Planning

Inventory Analysis

Customer ServicesCustomer interface

Order Handling

Customer Allocation

Consignment/VMI Inventories

LogisticsDie Bank/RDC/FGI handling

Kitting

Shipments

Impex

Program ManagementApple Demand / Supply Planning

SUPPLY

Supply Chain ExcellenceS&OP Process Improvement

Reporting requirements

System requirements

Order

Scheduling

Binning

Officer

Forecast /

Demand

collection


Figure 15 stakeholders in the value chain

Stakeholders in transport are service oriented with limited or no influence on product capacity.

Important for supply to know are the lead-times and variance for calculation purposes.

In the LED assembly & test stage many stakeholders are present. A supply chain planning model will

bring more clearness about product availability, throughput time and lead-time variance. On the other

hand if there is no capacity and/or flexibility in the chain the variance and lead-time should be covered

by stock. The balance between flexibility and delivery reliability is one of the key items.

3.3.3 Methodology of stakeholder analysis

The implementation of an optimal solution depends on the stakeholders in relation with the problem

situation. For that reason it is important to be aware of the perceptions, personal goals and interests of

the stakeholders in an early stage. Besides the identification of the stakeholders, the stakeholders can

bring relevant information about our demand. Also the powers and dependence of the different actors

related to the discussed subjects are analyzed. The result of this analyze contributes to finding an

accepted and most suitable solution for both, the component supplier and the assembly and test

manufacturer.

3.3.4 The stakeholders’ action field

Based on the business goals and the functions involved it can be concluded that commitment exist for

the implementation of a supply chain planning model. Main barriers are uncertainties about the effects

of the supply chain optimization model on their daily job and the interaction between the several

involved stakeholders.

Component production

TransportLED

assembly & test

Component suppliers

Production planner

Production

manager

Engineering

manager

Development

manager

Product manager

Shipping:

Parcel /Express carrier

Logistics:

Logistics department

component supplier

Component planner

Production planner

Supply manager

Planning manager

Supply chain excellence

manager

Assembly & test

Production manager


An overview of the actors and their position is presented in Figure 16 . The figure divides the

stakeholders in four groups. The first distinction is based on the dedication of the stakeholder to the

implementation of a supply chain optimization model. Those stakeholders that are dedicated to build a

model are positioned on the positive vertical axis, those who oppose it are on the negative side of the

vertical axis. The second distinction is based on the critically of the stakeholders in the process. If the

participation of stakeholders is critical, read if their participation is required for setting up a supply chain

optimization model, they are positioned accordingly on the horizontal axis.

It is important to note that this figure gives a static overview. Stakeholders can switch positions and

change from non-dedicated actors when perceptions change or compensation is offered.

Figure 16 stakeholder diagram

3.3.5 Importance of stakeholders

As indicated in Figure 16, 11 stakeholders are directly involved in supply chain optimization. To avoid

problems due to stakeholders during the research it is obvious to manage the stakeholders closely by

organizing workshops and bi-lateral meetings during the full process. During the planned workshops the

planners learn the structure of the models and have the opportunity to ask questions and come up with

possible additional requests.

Some opposition for the planning model can be expected from the logistic department at the

component supplier (LCS). The planner of the LCS is currently taking ownership about the planning and

ordering process between the two locations. During the research the current jobs of both, the LCS

planner and the planner at the manufacturer side will be challenged. The planners probably will show

their use and try to counter against the supply chain optimum. Besides the overall optimum there is a

CriticalNon-Critical

Dedicated

Non-Dedicated

Production planner

Maarheeze

Production

manager

Maarheeze

Engineering

manager

Maarheeze

Development

manager

Maarheeze

Product manager

MaarheezeSupply manager

PenangProduction planner

Penang

Component planner

Penang

Production

manager Penang

Supply chain

excellence

manager

Planning manager

Penang


continuous drive to go for local optimum solutions. The success of the supply chain optimum solution

depends on in which degree these people can be convinced about the robustness, flexibility and trust

about the overall optimum. The main difference between the overall optimum and a local optimum

solution is that the overall optimum is not touchable for the people working at the supply side. At the

other side, the people working at the supply side are responsible for the overall performance. For this

reason development of supply chain performance indicators are in scope of this thesis.

Effective structure requires an internal consistency among the design parameters and contingency

factors (H. Mintzberg, 1983). During the preparation phase of this assignment as assigned by the Supply

Chain manager at head-office the Supply chain organization was looked-over. The Lighting organization

has created a new business unit called Lumileds. Lumileds is smaller and less old than the traditional

Lighting organization. Formalization of systems, processes, tasks profiles and specialized job functions

were less developed in Lumileds compare to Lighting. The strategic drive into more Supply chain

controls in the Lumileds business is from that perspective not that strange. Within the background of a

less formalized organization the system model requirements were drawn up.

The actor contributes to the list of requirements, poses constraints, recognizes performance criteria and

identifies areas of interest for research options; these are presented in Table 1.

The actors as described in table 1 are all part of the Lumileds organization but part of different local

orientated organizations. All main stakeholders have a functional relationship with the Lumileds global

supply chain manager, the principal of the assignment to improve the supply chain controls. Within the

vision of the global supply chain manager supply chain controls are an enabler to jump to a next level of

maturity, meaning: reliability improvement, throughput-time reduction, stock reduction and an

improved customer service level.

The production planner of Maarheeze (the component supplier) is the current process owner and takes

care about the current stock positions at the manufacturer. The production planner Maarheeze has a

key role in the in the current way of working within Lumileds. According the production planner of

Maarheeze the expected improvements can be managed by the currently available planning tools.

The component planner Penang is responsible for managing the component supplier and the

components goods flow at the manufacturer. The manufacturer has direct contact with the final

customer in the total chain, for that customer satisfaction is one of the key drivers of the overall

organization.

The supply chain excellence manager is responsible for supply chain improvement projects within

Lumileds, for that reason the SC excellence manager is an ambassador and a sponsor for our

improvement project.

Finally, the material requirement manager has the responsibility for the overall incoming goods flow.

The component planner Penang is reporting to the requirement manager. The material requirement


manager has a need for supply chain performance reporting and model decision parameters to set

which influences performance indicators.

The main stakeholders as mentioned in table 1 have different interests in the results of the supposed

improved situation. At the one hand the production planner of the component manufacturer has his

own way of working and has limited interests to change the current situation. At the other hand the

manufacturer is facing an unreliable supplier as key component supplier. In the mindset of the Lumileds

an improved set of supply chain controls are necessary to manage the fast growing market with all the

expected product proliferations. The global supply chain manager asked the supply chain improvement

manager and the material manager to take care about the success of the supply chain control

improvement project. The planners at the manufacturer and at the component supplier are both key

knowledge owners necessary to design a most suitable supply chain control model.


Table 1 requirements, criteria, constraints & design option stakeholders

Main

stakeholder

Requirements Criteria Constraints Design options

Production

planner

Maarheeze

The interrelation

between

departments in the

value chain should

be efficient

Minimal

transaction time

Efficiency in

process design

Ownership of

optimization model

and decision making

process.

Vertical supply

chain integration

Component

planner Penang

The risks should be

covered by stock

policies.

Capacity

constraints to be

covered in the

S&OP

No below customer

service level

Options for lead-

time, variance

and service level

regulation.

Supply chain

excellence

manager

The model should be

stable and

predictable

To be ready within

4 months

- Options for lead-

time, variance

and service level

regulation.

Design of

performance

measurements

Material

manager Penang

The model should be

able to simulate

scenarios’

- - Design of report

generation

To resolve these conflicts effectively and turn the supply chain into a weapon for gaining competitive

advantage requires the development of an integrated supply chain driven by the needs of the business

(G.C. Stevens, 2007).

3.4 Conclusion analysis The need to focus more on supply chain planning and optimization instead of local optimization plans

has ascended the past few years. The LED market is growing enormously and actions to save the

company market share are necessary to survive. Industrial customers can switch, can go for a more

traditional solution or something don’t know yet. All these different kinds of behavior provoke business

losses and will harm continuity. In case industrial consumers decide not to purchase, this costs Philips

billons of turnover per month. However there are some appropriate ways to determine an optimal

solution for the optimization challenge. Research has indicated that are required actions to optimize the

LED supply chain. Without actions there will be too high inventory levels, utilization- and yield losses


with resulting financial losses. Current losses are caused by a lack of communication between the

component supplier and the manufacturer. Figure 11 shows a delivery mix performance (=CLIP;

customer line item performance) a score of 40% (average) with exceptions to below 20%. This mix

performance results in an additional uncertainty in the Lumileds supply chain of 5%.

Sub research question 3: which processes are related to the supply chain planning processes and what

are the independencies, is answered within chapter 3. As discussed in chapter three one of the enablers

of an optimal supply chain are communication and clear roles and responsibilities related to the supply

chain planning processes. The conclusions made after analyzing the lumileds supply chain is that the

current task and responsibilities are not sound and clear communicated throughout the Lumileds supply

chain including the key component suppliers.


Design and modeling phase

The analysis phase illustrates the need for Philips Lighting Lumileds to focus on growth. To make this

possible, the controls in the supply chain must be optimized. Information from the assembly and test

manufacturer and the component supplier can be used to identify the supply chain uncertainties and

can be turned into safety stock level, scheduling and replenishment advice. In the design phase, the

process and model to draw up a supply chain control advice will be designed. The basic building blocks

for the model are the involved product families and their uncertainty. Finally, an improved planning

method for the component supplier will be introduced.


4 Analysis of component through-put-time

4.1 Introduction Figure 17 is a simplification of the complex supply chain of the component supplier and the assembly

and test manufacturer. The 13 process steps in the production of the component are divided into two

stages; a front- and a back-end. The customer decoupling point is arranged via a supermarket between

the front- and back-end.

Figure 17 Simplified Supply chain of Lumileds

Managing stock levels asked for visibility in throughput time and for the possible uncertainties in the

chain. Section 5.2 and 5.3 elaborate about the difference between the front- and back-end processes at

the component supplier. The differences in throughput time per product family are analyzed based on

literature (normality testing) with Minitab and insights of professionals in the field (section 5.4).

For a better understanding of the component supply chain and the complexity at the component

supplier see below figures (Figure 19, Figure 20, Figure 21 and Figure 22). In the first picture the front-

end of the component supplier is visualized. Starting with the raw materials, 5 process steps end in a

supermarket. The semi finished product is stored in a supermarket to shorten the customer reaction

time as much as possible. Figure 22b shows the total throughput-time of the Front-end of the

component supplier. Figure 18, shows the distribution of the throughput-time of the front-end

processes. In this picture the distribution of the front-end is challenged between a normal distribution, a

3 parameter Weibull distribution, a gamma distribution and a 3-parameter gamma distribution. The

probability plot as shown in figure 18 measures how well the data follow a particular distribution. The

better the distribution fits the data, the smaller this AD (Anderson-Darling) statistic will be. When trying

to determine which distribution the data follows multiple Anderson-Darling statistics can be applied to

compare the distributions. The distribution with the smallest Anderson-Darling statistic has the closest

fit to the data. For our data set of the front-end this will mean that a 3-parameter gamma distribution

fits best. However, the data is selected in a period without any improvement action implemented like

fifo at the work centers and buffers in between. Due to the limited process controls production batches

are held up with or without any reason resulting in longer and unreliable throughput-times.


Figure 18 Probability plot front-end component supplier

Figure 19 Front-end component supplier

The strict fifo lanes between the process steps in the front-end are buffers for compensating the

different processing times. The fifo lanes and process steps are supervised by a MES (manufacturing

execution system), based on system setup production batches are managed through the line.

The back-end of the component supplier has 7 or 9 process steps. Product group A has 7 process steps

and product group B has 9 steps. For product group B, one process step is executed at a different

location (outsourced to a subcontractor).

7550250

99.99

99

90

50

10

1

0.01

T hroughput-time in days

Pe

rce

nt

100.0010.001.000.100.01

99.99

90

50

10

1

0.01

T hroughput-time - T hreshold

Pe

rce

nt

100101

99.99

99

90

50

10

1

0.01

T hroughput-time in days

Pe

rce

nt

100.010.01.00.1

99.99

99

90

50

10

1

0.01

T hroughput-time - T hreshold

Pe

rce

nt

Gamma

A D = 13.636

P-V alue < 0.005

3-Parameter Gamma

A D = 3.869

P-V alue = *

Goodness of F it Test

Normal

A D = 44.149

P-V alue < 0.005

3-Parameter Weibull

A D = 8.791

P-V alue < 0.005

Probability Plot for front-end component supplier

Normal - 95% C I 3-Parameter Weibull - 95% C I

Gamma - 95% C I 3-Parameter Gamma - 95% C I

Granulation Pre-grindingSinteringBBOPressing

FIFO

Max X

FIFO

Max X

FIFO

Max X

FIFO

Max X

Raws

Front-end

907560453015

400

300

200

100

0

Days

Fre

qu

en

cy

Throughput-time front-end component supplier


Figure 20 Back-end component supplier

The various throughput-times for product group A and B are shown in Figure 21.

Figure 21 Throughput-time back-end component supplier

The level of uncertainty is visualized in the standard deviation in Figure 21, for A and B types products is

such unreliable that extra stock at the assembly and test manufacturer are necessary. Reducing the

variation in the back-end of the component supplier needs several actions to succeed. Important is that

the model should be able to plan with all uncertainties as currently available. Optimization of the

uncertainties at the component supplier is not a primary task of the solution design but the model

should also be useful to show the effects via the stock levels due to the high level of variance in the

chain.

Dichroic

Aachen

Coating

Measurem

ent platelet

100%

Visual

inspection

100%

SeparationSeptaping

Waf

measuringGrindingGrind-taping

FIFO

Max X

FIFO

Max X

FIFO

Max X

FIFO

Max X

FIFO

Max X

FIFO

Max X

FIFO

Max X

WW

Hikari

Back-end

75604530150

300

250

200

150

100

50

0

Throughput-time in days

Fre

qu

en

cy

Throughput-time back-end component supplier A-type

6048362412

14

12

10

8

6

4

2

0

Throughput-time in days

Fre

qu

en

cy

Throughput-time back-end component supplier B-type


Figure 22 Entry process steps assembly & test manufacturer

Figure 22 shows first steps of the assembly and test at the manufacturer in the Asia-pacific region.

Interesting is the second incoming goods flow from another key-component supplier located in

Singapore. Demand planning information of the second key-supplier will be used as input for the weekly

replenishment orders at the component supplier.

4.2 Front-end In Figure 23 the cycle time of the front-end is shown. At the primary X-axis the median cycle time in days

is mentioned. At the secondary X-axis the WIP is indicated as the total number of different batches in

the line during the time frame as mentioned at the Y-axis. The vertical bars in the graph representing the

variance in the selected period. Remarkable is the difference in spread over the selected period. The

selected period is chosen based on the following pillars: 1. Ramp-up of the based products was started

at the beginning of quarter 2 of 2010, the period before can be identified as a development phase. 2.

Medium quarter 3 2010 the first effects of the model implementation are visible in the weekly

performance measurement. Based on the analysis as presented in Figure 23 the conclusion is that the

variance of the cycle time in the front-end process is growing. At the other hand (and in line with the

first conclusion) the work in process (WIP) is increased too. Due to limited process control the cycle time

of the production batches will be influenced. The median of the cycle-time is relative stable with a

growing variance and WIP. Due to limited process control the cycle time of the production batches can

be influenced in a positive or negative way. More details about the sub processes of the front-end part

of the component manufacturer are described in appendix D.

Figure 23 Median throughput time front-end

Set matching

(combine tile &

platelet)PnP sorting

Platelets in

DieBank

I

Platelets in

DieBank

I

LLM 2702 RAW

I2702 DB

Singapore

Saber Fab

Tile Altilon & Flash

Tile WW & Amber FIFO

Max X

I

2702 DB

Flash

Visual

inspection

100%

Expand to

metal foton ring UV Cure Quality control

100%

Lumi visual

inspection

100%

0

10

20

30

40

50

60

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

14 16 18 20 22 24 26 28 30 32

WIP

Med

ian

Cycle

Tim

e (d

ays)

Work Week 2010

Cycle-time front-end component supplier

CT (days) WIP


A storage location is located between the front- and the back- end. Based on stock keeping rules as

described in section 5.5 an optimal service level towards the back-end is managed.

4.3 Back-end The variance of the back-end processes is representing by the vertical error bars in Figure 24. The back-

end process has a direct link with the assembly and test manufacturer. Based on the production process

of the manufacturer the back-end of the component supplier is managed. Due to uncertainties in the

production process of the component supplier extra stock is necessary at the manufacturer to cover the

risks of the uncertainties. Comparable with the front-end no strict hand over rules from one process

step to the other are used, the limited level of controls ends in a higher level of uncertainties.

Figure 24 Median throughput time back-end

More details about the sub processes of the back-end part of the component manufacturer are

described in appendix E.

Variance in the back-end of the component supplier is caused by several topics. Most of the struggles

responsible for the current variance in lead-time are related to the strong ramp-up at the production in

2010. Based on a defect analysis (same period as Figure 23) a Pareto analysis was prepared, see Figure

25.

0

10

20

30

40

50

60

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

14 16 18 20 22 24 26 28 30 32

WIP

Med

ian

Cycle

Tim

e (d

ays)

Work Week 2010

Cycle-time back-end component supplier

CT (days) WIP


Figure 25 Pareto negative effects on throughput-time component supplier

Figure 25 shows the main sources of the variance as present in the factory of the component supplier.

Important in the Pareto analysis is that five out of nine key issues are topics related to staffing and/or

the organization. Related to Mintzberg the organization is changed from an adhocracy towards a

Machine Bureaucracy. Process alignment is moved from mutual understanding towards a Standard

Operating Procedure (SOP) in which the work processes are described including controls, escalation

paths, communication flow and performance measurements.

The Pareto as presented in Figure 25 is a result of a measurement as held in week 10 till week 40 2010.

The negative effects of the defects were not registered only the root-cause of the incidents, for that

reason no detailed overview of the total delay can be presented.

4.4 Supermarket stock level calculation In the previous sections the process of the component supplier is discussed. Optimizing the supply chain

with a focus on short lead-times to the LED assembly and test manufacturer are the most important

design objectives. Meaning, the supply from the component supplier to the manufacturer should be

boring predictable. A common way to determine the number of Kanban tickets is described with the

formula (Marmostein and Zinn, 1990):

[5]

KB = number of kanban tickets

D = forecast demand (weekly)

LT = production lead-time back-end component supplier (in weeks)

Z = service level

STD = standard deviation of demand * coefficient of variance

Q = batch size

Sta

ffin

g c

ap

acity is

sue

Cap

acity a

lignm

ent

sub

co

ntr

acte

rs

Eq

uip

em

ent d

ow

n

Qualit

y is

sues

Main

tenance -

pla

nned

Main

tenance -

unp

lanned

Prio

ritiza

tio

n o

f activitie

s a

t shif

t le

vel

Fix

ed

bre

ak-t

imes

Deta

iled

activity s

ched

ulin

g

mis

sin

g

Sup

ply

co

nsum

ab

les

Share

d r

eso

urc

es w

ith o

ther

pro

ductio

n f

acili

ty

Sup

po

rtin

g m

ate

rials

are

m

issin

g

Wro

ng

measure

ment

0%

20%

40%

60%

80%

100%

0

5

10

15

20

25

30

35

40

45

Cum

ula

tive %

De

fec

ts

Causes

Pareto analysis defects production component supplier

Vital Few Useful Many Cumulative% Cut Off % [42]


The number of kanban tickets per product type is determined by the demand during lead-time plus the

service level multiplied with the standard deviation of the demand. The service level is a dimension free

number and is represented by the value Z and can be calculated in excel (NORMSINV), see Figure 26.

Figure 26 Z value development

This formula to calculate the number of Kanban tickets in the front-end of the component supplier and

especially the Z-factor calculation presumes that the lead time and demand have a normal distribution.

The demand pattern as discussed in section 4.1 resulted in a 3-parameter Gamma distribution as best fit

but due to the limited controls as currently available resulted in many outliners in the data set. The data

as used in the analysis will not reflect the goods flow after implementing the controls as discussed;

therefore a normal demand distribution is assumed.

The model generates advice for the safety stock and the total number of Kanban tickets (= safety stock +

WIP) for a desired service degree to the back-end based on the demand (is S&OP) of the back-end. The

lead-time and the coefficient of variation are determined for the different product groups. The demand

pattern is unique for all products. The model combines the lead time and the demand pattern and uses

a Z value as service level indication, in this model 95%, 97.5% and 99%.

The results of the model are represented in Appendix F. These outcomes presume that it’s possible for

the front-end to serve the back-end based on daily scheduling /replenishment and a monthly review

process to align the demand uncertainties as used in the calculation (N-2) and the variance as it was in

practice during the last period (N-1). The average results for the data set are presented in Table 2.

0

0.5

1

1.5

2

2.5

3

0.9 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 0.995

Serv

ice

fac

tor

valu

e

Service factor (Z)

Z value development


Table 2 Average stock level advice for data set

Dividing the safety stock production batches by the average production capacity of the back-end per day

gives the safety stock level in days. The Work-In-Process (WIP) based on the advice model is 106 Kanban

tickets. This part of the calculation is stable for all service levels due the fact that the WIP covers the

expected demand; the safety stock covers the uncertainties and the demand during lead-time. The total

stock (total number of Kanban tickets) is the sum of the safety stock and the WIP.

95% 97.50% 99%

Safety stock (tickets) 52 58 66

Safety stock (days) 5.2 5.8 6.6

WIP (tickets) 106 106 106

WIP (days) 10.6 10.6 10.6

Total stock (tickets) 158 164 172

Total stock (days) 15.8 16.4 17.2

Service level


5 Design setup

5.1 Introduction This chapter presents the design setup for the model which optimizes product availability based on a

product family specific variance factor with attention to stock, service level and lead time. In the current

situation all SKU’s are treated similarly. Therefore an improvement opportunity is to adapt stock levels

per product family to fit better with the specific SKU family identities. The design of the model is

explained hereafter, starting with the supply chain performance indicators (section 4.2). Section 4.3

discusses the end-to-end risks in the supply chain and how they are related to the model. Section 4.4

elaborates on the input data for the design. Finally, section 4.5 discusses design setup conclusions.

5.2 Performance measurements Performance management (as discussed by W. Jammernegg ao in: Performance improvement of supply

chain processes by coordinated inventory and capacity management) illustrates how successful the

uncertainties are managed. All activities that are done, should improve service to our customers.

Increased performance of the component supplier has a positive effect on service and will therefore

potentially result in more sales and profit. Improved product availability through better balanced (to

cover the uncertainties) stock positions lead to improved end-to-end delivery reliability, expressed in

CLIP (Customer Line Item Performance) and CVP (Confirmed Volume Performance). The appendix on

page 79 shows the relation between stock levels on the one hand and CLIP and CVP on the other hand.

5.3 Uncertainty The relation between uncertainty within the supply chain, service and stock levels is represented in a

causal model (Figure 27).

Figure 27 Causal model of uncertainty and customer performance

CLIP

Stock levels at

manufacturer

+

Uncertainties

product

families-

Stock levels at

component

supplier

+

Responsivene

ss on demand

+

Service level

Demand

during lead-

time

Demand

variance

+

+

+

S&OP

reliability

+

Customer

demand

+

-

Production

lead-time

Service level

Supply

variance

+

+

+


The component production and the related uncertainty are positioned upstream in the supply chain.

When these uncertainties increase, the delivery performance potentially decreases. As discussed, stock

levels influence the delivery performance to the assembly and test manufacturer. The stock levels both

at the component supplier and the assembly and test manufacturer are affected by desired service level,

lead time and (supply or demand) variance.

5.3.1 Replenishment strategy

A proper selection of a replenishment strategy is one of the keys to achieving low inventory while

maintaining high customer delivery performance (P. Suwanruji eo, Evaluating the effects of capacity

constraints and demand patterns on supply replenishment strategies, 2005). In Table 3 a comparison is

made between three replenishment strategies; using DRP /MRP (Distribution /Material Requirement

Planning), ROP (Re-Order-Point) and KBN (Kanban) were compared.

Using DRP/MRP strategy requires full visibility of order and inventory information across all locations in

the Supply Chain with periodic review. ROP is a reactive strategy with replenishment decisions based on

continuous review. KBN is a reactive replenishment strategy; inventory within a replenishment loop is

controlled by a fix number of cards. Kanban cards in a replenishment loop were allocated to inventory at

a location, inventory in transit to a location and unfilled orders placed to an upstream location.

The three replenishment strategies are compared under two levels of Manufacturing Constraints (MC)

No time-delay capacity constraint and Time-delay capacity constraint. No time-delay capacity constraint

ignores time delay operations by setting all setup and part processing times equal to zero

(manufacturing resources have no capacity constraint). The time-delay capacity constraint, production

resources at manufacturing are able to process one order at the time. A capacity constraint exists and

batches waiting to be processed must join the queue.

The second level of challenge is found in the Demand pattern (DP): level demand and seasonal demand.

During a level demand pattern the expected period demand was assumed to be stable through time for

each item. For the seasonal pattern, the demand pattern is assumed to follow a sinusoidal pattern with

a cycle length equal to one year.


Manufacturing

constraint (MC)

Demand pattern

(DP)

Replenishment strategy (strgy)

DRP/MRP (strgy 1) ROP (strgy 2) KBN (strgy 3)

No capacity

constraint

Level demand

Responds slow to

changes introduced

by demand

uncertainty.

Responds

directly to

changes by

demand &

supply

uncertainties.

Results in

steadiest stream

of arrivals which

in turn results in

shorter average

queues.

Seasonal demand

Responds on

demand variation

including a season

correction based

on history and

forecast input.

Responds on

demand

variation

forecasted for

the next period.

Does not use

forecasting

information and

cannot utilize

backorders

information.

Capacity

constraint

Level demand

Effective at co-

coordinating

material flow and

limited waiting

time for the same

level of component

inventory

Score between

KBN and MRP –

managing the

material flow

and anticipates

on the demand

uncertainties.

Beneficial when

demand

uncertainty is of a

random nature,

results in

relatively low

inventory in front

of capacitated

resources

Seasonal demand

Responds on

demand variation

including a season

correction based

on history and

forecast input,

capacity limitations

are integrated.

Backorder

information is

taken into

account in the

next calculation

but frequency is

lower than at

the MRP

strategy.

No backorder

information is

maintained.

Table 3 Replenishment strategies (P. Suwanruji ao, 2005)

The green framework indicates the generic best suitable replenishment strategy per MC & DP

combination.


Key characteristics of the Lumileds Supply Chain are directly linked to extreme business growth and the

many uncertainties in the current Supply Chain. Seasonal effects are not visible at the current state,

meaning no bullwhip effects of seasonal demand in the Supply Chain. This implies that for Philips

Lumileds ROP or Kanban replenishment strategy is most suitable for the Lumileds Supply Chain

(depending on the existence of capacity constraint).

To generate a stock advice, the advice model is divided in three sub models: the Kanban component

supplier, the economical stock value calculation and the replenishment model (Figure 28).

Figure 28 Concept advice model

5.3.2 Replenishment strategy component supplier

The upstream part of the chain is managed by kanban principle. Based on the main drivers as

mentioned in the box in Figure 17, the number of Kanban tickets is determined. Per product family the

value of the main drivers can be different. The performance of the Kanban concept is managed by daily,

weekly and monthly processes. Each period has a different focus; the daily process has a focus on

starting the right number of Kanban tickets per product. The weekly process has a focus on balancing

throughput time in relation with customer demand. The last process (the monthly) focuses on updating

supply variance, lead time, batch sizes and service levels. The uncertainty to manage is covered by the

supply variance, in the variance we have two parts lead-time and batch size. The different products are

divided in groups (product family, see appendix A) based on their uncertainty profile.

5.3.3 Re-order-point calculation

Calculating the re-order-point (ROP) is a monthly process. The S&OP update brings new input about

customer’s behavior. The ROP level consists out of two parts; the first part covers the demand during

lead time the second part covers the safety stock (see also par 2.5 Supply chain control models). The

uncertainty to manage is formed by the demand pattern as represented by the real shipments. The

demand and supply uncertainties are input to determine the necessary safety stock in the advice model.

Re-order-point

calculation

Replenishment

orders

Stock

levels

Manuf

acturer+/- +/-

S&OP

Supply variance

Demand variance

Moving average

Lead-time

Service level

Product life-cycle

phase

Stock take

WIP

In-Transit

ROP

Kanban

component

supplier

S&OP

Supply variance

Lead-time

Batch size

Service level

+/-

+/-


5.3.4 Replenishment orders

Based on the boundaries as set in the ROP, the replenishment orders from the manufacturer to the

component supplier2 are calculated. The products on stock at the manufacturer are in-transit and in the

work-in-process are taken into account and are a minus on the calculated demand. The calculated

demand is based on a process step at component supplier1 of the assembly and test manufacturer. The

result of this action is a decrease of one week of stock to cover (see also Figure 29).

Week 1 Week 2 Week 3 Week 4

Assembly & test

manufacturer

Component supplier 1


Assembly & test

PP

PP

Production &

transport

Production & transport

PP = production plan

Difference in production lead

time to cover with stock

Figure 29 Planning of replenishment orders

5.4 Data The data used for this research is taken from the MES (Manufacturing Excellence System) of the

component supplier, the ERP system of the assembly and test manufacturer and the sales and

operations planning of the business unit Lumileds. The selected data represents the situation before the

results of the proposed solution were visible in the supply chain performance indicators of Lumileds.

4.4.1 MES

Data from the MES database is filtered on product family. The lead-time per process step is calculated

by subtracting all delivered production batches from the database. In the standard MES report the

throughput time per process element is measured in hours, the totals are expressed in days of

throughput time. All delivered batches (orders) in the selected period are taken into account no

exceptions. The data integrity for all figures is ensured by using production (MES) data integrated in the

ERP of Philips Electronics N.V.

5.4.2 ERP

Required data for calculating the ROP are the shipped products out of the assembly and test

manufacturer. Based on the shipped out data the demand variation is calculated. In the weekly

replenishment calculation a stock take out of the ERP system is copied into the model. The integrity of

the data is safeguarded by Finance and accounting of the assembly and test manufacturer; all system

mutations have financial consequences and for that controlled by F&A and periodically checked by an

independent accountant.


5.4.3 S&OP

Customer demand is consolidated in an S&OP for the Lumileds business. A monthly update provides the

business new insight in the customer’s forecast with an outlook for 15 months. The supply chain

department of the business unit Lumileds is responsible for the content and reliability to translate sales

data into a demand plan. Based on the demand plan the supply plan is generated, a part of the overall

supply plan is for the component supplier. This component supply plan is the S&OP input for the

component supplier. The component supplier will use the S&OP as input for the day to day business and

on the other hand for mid- and long term capacity increase/decrease business decisions.

5.5 Conclusions Improving service through the Lumileds supply chain is the key objective of this research. Optimizing the

stock levels has a direct and positive effect on the customer service (CLIP). When the stock levels are

better adapted to the end-to-end uncertainties in the supply chain, the CLIP will improve because more

risks or uncertainties are covered.

The supply chain is represented as an uncertainty model. There are three types of uncertainties

influencing stock safety levels: at the component supplier lead-time and batch size variances and at the

assembly and test manufacturer the variance in customer demand. All those uncertainties including the

lead-time are influencing required stock levels at the component supplier and at the assembly and test

manufacturer. A lead-time means a higher stock level to handle the demand during lead-time to

guarantee delivery performance. Reducing lead time will directly effect in lower required stock levels.

Chapter four (partly) answered sub question 2 (which planning model is most suitable for the supply

chain).


6 Model design

6.1 Introduction model design In the previous chapters the design objectives and boundaries for the model are discussed in depth. In

the next section the structure of the model, based on literature, is explained and there is some attention

to the relation between the data and the choices of the model structure. The chapter ends with a

description of an excel tool which will be used to determine the optimal Re-Order-Point and safety stock

levels. This tool is developed together with field experts to ensure the practical workability of the tool

and the acceptation of the end users.

6.2 Evaluation design objective First point of attention is to challenge the accuracy of the specified requirements, the quality of the

model and the acceptance during the hand-over to operations at the end of the project depends greatly

on the investment during this activity (Lindland et al., 1994). The first step in the design phase is the

setup of a conceptual model. This is taken in section 4.3. The system and its environment are identified

and the mutual relations are determined (Figure 27). The relations are translated into a conceptual

model, wherein the three sub models are represented (Figure 28). In the analysis phase the design

objectives are identified.

1. Design a model to optimize the stock positions throughout the Lumileds Supply Chain, based on

product specific Supply chain uncertainties.

2. Design a model to improve the service level of the Lumileds Supply Chain.

The advice model has a focus on these two objectives and the success depends strongly on the degree

of fulfillment of these two objectives.

6.3 Model design based on the theory As described in Figure 28 the overall solution is designed in 3 sub models; one model for the Re-Order-

Point calculation, a model for replenishment determination and a Kanban model for balancing the

production of the component supplier (chapter 5).

6.4 Practical design and model result Based on the inputs of chapter 4 (design setup) a model is designed as a pilot. This first design was built

based on the requirements gathered during interviews and workshops. After evaluation a next version

of the models were created. Finally, with input of field experts the models and the results out of the

pilot were created. Figure 30 shows the first level of the IDEF0 model of the designed situation. Twelve

input data flows, six control flows, one supporting mechanism and one output dataflow are the number

of instruments used in the models. In the next paragraphs the design of the models is discussed based

on the IDEF0 model.


Figure 30 Model design (IDEF0 level 0)

Figure 31 shows the supply chain control model of the component supplier. The model generates advice

for several topics needed at the different stages in the chain. The final result of the integrated models is

a consolidated stock data overview based on re-order-point levels. The re-order-point calculation is a

monthly process based on the boundaries as set during the monthly cycle. The replenishment orders will

be generated in the weekly cycle. The Kanban calculation of the front-end of the component supplier

takes the demand forecast out of the re-order-point calculation.

In the next sections the model is designed and built to satisfy the design objectives within the design

boundaries.

Supply chain planning model

A0

Pe

riod

MA

T

Pro

d. L

ife C

ycle

Se

rvic

e le

ve

l

Ca

p. c

on

stra

ints

Ca

rrier le

ad-tim

e

S&OP

Shipments

Lead-time platelets

Prod. Fam. table

SM comp supl.

GR prod order

GGI starts

Yield

Stock overview

In-transit

Platelet distribution

WIP comp. Supl.

ME

S

Consolidated stock

overview


Figure 31 Model design (IDEF0 level 1)

6.4.1 Monthly ROP calculation

The monthly re-order-point calculation (A1 in Figure 31) generates the new or updated Re-order-point

values. The inputs for calculating the re-order-point levels are the platelet distribution, the S&OP, the

actual shipments of the last 60 days, the lead-time of the back-end of the component supplier including

the carrier lead-time and the product family table. In the product family table the decision is taken if a

product is able to move over from a manual to an automated stock calculation method. The coefficient

of variation (CV) is the decision making qualifier. If the CV value is above 1; manual intervention is

necessary in the economical stock level calculation. If the value is below 1; no manual intervention is

required, the model will take care about the uncertainties in the supply chain. There are several ways to

determine the ROP level (as partly discussed in chapter 2.5). The most common, simple way to

determine the re-order-point level is described with the formula (Charles Atkinsin, 2005):

[6]

ROP = Re-Order-Point

DEM = Demand (weekly)

LT1 = Median lead-time back-end component supplier including transport time

Z = Service level

LT2 = Median LT1 – lead time GGI/Saber


Dvar = Demand variation

Svar = Supply variation

The ROP is determined by two parts: the first part is for covering the demand during lead-time, the

second part is for covering the uncertainties in the chain (safety stock). The safety stock is determined

by the service factor Z. This number represent S the required service level and can be calculated in excel

(NORMSINV). Figure 32 shows the results of the Re-Order-Point calculation model.

The ROP (Re-Order-Point) is the sum of Demand during lead-time (Dem LT) and the Safety Stock. The

Demand during Lead-time is a calculation of the Weekly demand and LT1 (lead-time back-end

component supplier + transport lead-time + replenishment lead-time). LT2 is used in the safety stock

calculation and is used together with LT1, LT2 stands to the difference between the lead-time demand

and the GGI starts (gold to gold interconnect process: start production process marked as leading within

the SC of Lumileds). The supply variance as used in the safety stock calculation is based on the supply

standard deviation related to throughput-time and the supply standard deviation related to quantities.

The demand variance is calculated with data based on products shipped, with the assumption that the

demand variance of the past is comparable with the demand variance of the future. The Z value is

directly related to the service level as indicated in the last column. The yellow fields are manual input

possibilities and will be updated by the production planners of the component supplier and the

manufacturer. The product family as mentioned in figure 32 is the total product portfolio of the lumileds

component supplier. Figure 32 reflects the ROP calculation for January 2011.

Figure 32 ROP calculation model

6.4.2 Weekly replenishment

The weekly replenishment calculation (A2 in Figure 31) is calculated based on the boundaries as set in

A1 in Figure 31 (the ROP). The input data flows are needed to calculate the needs for the coming period.

A common and simple way to determine the replenishment levels is described with the formula:

ROP calculation sheet(MAT)

Product Line Bin ROP Dem LT. Safety stock Weekly dem. LT 1 LT 2 Suppl var. Dem var. Z Safety[wks] Avg stock [wks]Supply Std dev BESupply std dev # transit fileLT BE LT GGI Service l.Product family A

330 479,536 307,717 171,819 137,199 2.24 1.00 0.48 0.59 1.64 1.25 1.75 0.46 0.15 1.24 1.71 95%

440 418,344 280,277 138,067 137,199 2.04 1.00 0.22 0.59 1.64 1.01 1.51 0.17 0.14 1.04 1.71 95%

Product family B

222 - - - - 2.43 1.00 0.74 0.48 1.64 0.72 0.18 1.43 1.71 95%

333 4,679,922 3,083,530 1,596,392 1,438,981 2.14 1.00 0.30 0.48 2.05 1.11 1.61 0.24 0.18 1.14 1.71 98%

444 - - - - 1.00 1.00 0.11 0.48 1.64 - 0.11 - 1.71 95%

555 10,988,393 7,338,801 3,649,592 3,357,621 2.19 1.00 0.31 0.48 2.05 1.09 1.59 0.21 0.22 1.19 1.71 98%

Product family C

241 39,462 23,018 16,444 6,975 3.30 1.59 0.52 0.98 1.64 2.36 2.86 0.46 0.24 2.30 1.71 95%

242 39,462 23,018 16,444 6,975 3.30 1.59 0.52 0.98 1.64 2.36 2.86 0.46 0.24 2.30 1.71 95%

Product family D

111 42,374 25,958 16,416 8,259 3.14 1.43 0.15 1.00 1.64 1.99 2.49 0.12 0.09 2.14 1.71 95%

222 2,491,901 1,275,450 1,216,451 368,932 3.46 1.74 0.40 1.45 1.64 3.30 3.80 0.34 0.22 2.46 1.71 95%

333 - - - - 3.61 1.90 1.20 1.45 1.64 1.20 0.10 2.61 1.71 95%

444 104,250 63,865 40,385 20,601 3.10 1.39 0.23 1.00 1.64 1.96 2.46 0.13 0.19 2.10 1.71 95%

555 1,614,924 1,096,981 517,943 298,789 3.67 1.96 0.23 0.74 1.64 1.73 2.23 0.09 0.21 2.67 1.71 95%

666 - - - - 3.63 1.91 0.41 0.74 1.64 0.40 0.10 2.63 1.71 95%

777 226,717 168,586 58,131 64,841 2.60 1.00 0.19 0.52 1.64 0.90 1.40 0.18 0.06 1.60 1.71 95%

Product family E

1820 - - - - 2.48 1.00 0.85 0.55 1.64 0.37 0.77 1.48 1.71 95%

1813 325,994 235,151 90,843 96,430 2.44 1.00 0.58 0.55 1.64 0.94 1.44 0.17 0.55 1.44 1.71 95%

1806 1,685,156 1,135,556 549,599 341,888 3.32 1.61 0.59 0.55 1.64 1.61 2.11 0.43 0.40 2.32 1.71 95%

Product family F

999 1,229,283 575,327 653,956 231,720 2.48 1.00 0.56 1.68 1.64 2.82 3.32 0.35 0.44 1.48 1.71 95%

Product family G

444 569,550 334,922 234,628 81,688 4.10 2.39 0.41 1.00 1.64 2.87 3.37 0.34 0.22 3.10 1.71 95%

555 141,882 83,439 58,443 20,422 4.09 2.37 0.41 1.00 1.64 2.86 3.36 0.34 0.22 3.09 1.71 95%

Total 25,084,392 16,054,333 9,030,059 6,621,257


[7]

Repl = Replenishment order to the component supplier

ROP = Re-Order-Point level

WIP = Work-in-process back-end component supplier

InT = In-Transit between component supplier and the LED assembly and test manufacturer

Stock = Physical stock availability at the LED assembly and test manufacturer

Prod = Production as planned for the selected time period

Figure 33 shows the results of the replenishment calculation model for the last quarter of 2010 (the first

full quarter that the model is fully used in the Lumileds organization), the selected products are

reflecting the overall way of working.

Figure 33 Weekly replenishment calculation

Figure 33 shows the weekly replenishment quantities per product type. Per week Usage, Supply Pipeline

Economical stock, Economical stock end of week and on-hand end of week are calculated. The usage is

based on the production plan of another key supplier as discussed in 5.3.4 Replenishment orders, in

case no starts are planned at the GGI, the ‘Buildplan /S&OP’ will be taken as guideline. At the supply

side, ‘Planning Mhz’ indicates the replenishment order based on the calculation as discussed. The

‘confirmed plan Mhz’ are the quantities as planned at the component supplier. The difference between

‘planning Mhz’ and ‘Confirmed plan Mhz’ is related to extra information as available in the chain of

Lumileds (be. Expected extra sales not yet integrated in the S&OP or a possible temporary shutdown at

the component supplier start 2011). The ‘Pipeline Econ. Stock’ is a summary of the work in process at

the component supplier (back-end), ‘goods in transit’ and the components as available at the

manufacturer. The ‘Pipeline econ. Stock’ is summarized as ‘Econ Stock (qty)’ and weeks of forecasted

consumption. In case the ‘Econ Stock (wks)’ is lower than two weeks this fields will be red colored. The

21 Replenishment Calculation based on GGI Starts

Wknr 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052

Product family B, BIN 333

Backlog 0 0 0 0 0 0 0 0 0 0 0 0 0

Batchsize GGI Starts SGP 1,680,000 1,680,000 1,924,128 2,100,000 1,911,552 2,090,000 1,815,000 2,144,000 2,398,050 2,200,000 2,700,000 2,100,000 2,452,320

20960 Buildplan/S&OP GGI 1,512,000 1,512,000 1,731,715 1,890,000 1,720,397 1,881,000 1,633,500 1,929,600 2,158,245 1,980,000 2,430,000 1,890,000 2,207,088

90.0% Actual Pl/Attach

ROP 5,796,082 5,796,082 5,796,082 5,796,082 5,796,082 4,817,786 4,817,786 4,817,786 4,817,786 6,222,473 6,222,473 4,977,979 4,977,979

Batchsize Planning Mhz 1,416,000 1,062,000 944,000 944,000 236,000 0 0 0 0 0 0 0 0

118000 Confirmed plan Mhz 2,580,000 2,300,000 2,340,000 2,400,000 1,845,000 3,132,000 2,668,000 1,740,000 1,392,000 2,088,000 0 0 1,190,000

WIP Mhz 1,548,000 516,000 615,000 1,032,000 1,722,000 1,548,000 903,000 1,392,000 1,624,000 580,000 928,000 928,000 100,000

In-Transit 1,653,648 1,779,774 2,244,509 1,392,272 1,190,373 1,670,764 2,327,476 1,567,587 1,156,905 1,909,011 1,635,070 1,676,891 222,184

Stock take Penang 2,724,371 4,020,741 3,746,909 4,318,502 4,406,332 4,854,873 5,282,569 5,873,545 6,972,546 7,878,446 7,834,441 7,389,441 8,068,381

Econ. Stock (qty) 6,994,019 7,104,515 7,214,703 7,252,774 7,443,308 9,324,637 9,547,545 8,643,532 8,987,206 10,475,457 7,967,511 8,104,332 7,373,477

Econ. Stock (wks) 4.1 4.2 4.3 4.4 4.5 5.7 5.8 5.3 5.7 6.7 5.3 5.5 5.2

On-hand + InTransit 2,495,543 3,923,055 4,479,418 4,198,774 3,864,990 4,635,637 5,889,648 5,560,132 6,495,951 7,857,857 7,311,266 7,086,332 5,860,565

Product family B, BIN 555

Backlog 0 0 0 0 0 0 0 0 0 0 0 0 0

Batchsize GGI Starts SGP 1,120,000 1,120,000 1,282,752 1,400,000 1,400,000 1,482,000 1,320,000 2,600,000 2,930,950 2,500,000 2,700,000 2,100,000 2,431,360

20960 Buildplan/S&OP GGI 1,008,000 1,008,000 1,154,477 1,260,000 1,260,000 1,333,800 1,188,000 2,340,000 2,637,855 2,250,000 2,430,000 1,890,000 2,188,224

90.0% Actual Pl/Attach

ROP 3,932,191 3,932,191 3,932,191 3,932,191 3,932,191 3,554,717 3,554,717 3,554,717 3,554,717 6,473,904 6,473,904 7,768,685 7,768,685

Batchsize Planning Mhz 236,000 0 826,000 0 0 0 0 0 0 590,000 1,180,000 3,422,000 1,888,000

118000 Confirmed plan Mhz 1,677,000 1,700,000 1,710,000 2,200,000 1,368,000 1,710,000 1,710,000 2,166,000 2,736,000 2,280,000 2,808,000 2,574,000 2,925,000

WIP Mhz 0 1,677,000 615,000 903,000 1,161,000 1,032,000 1,419,000 1,140,000 1,140,000 2,052,000 1,856,000 812,000 1,972,000

In-Transit 2,222,499 81,398 1,241,114 1,590,318 1,223,946 2,127,451 511,366 1,143,207 1,082,338 1,224,475 1,973,875 2,231,481 2,153,156

Stock take Penang 2,554,185 3,185,424 2,491,348 3,711,629 4,260,958 3,952,499 4,886,614 4,461,075 4,280,424 4,871,116 3,931,361 3,228,353 3,962,697

Econ. Stock (qty) 5,445,684 5,635,822 4,902,985 7,144,947 6,753,904 7,488,150 7,338,980 6,570,282 6,600,907 8,177,591 8,139,236 6,955,834 8,824,629

Econ. Stock (wks) 2.4 2.4 2.0 2.8 2.5 2.7 2.6 2.3 2.3 2.8 2.8 2.3 2.9

On-hand + InTransit 3,678,573 2,011,838 2,724,462 4,293,947 4,330,427 4,819,950 4,137,980 4,270,482 4,174,762 3,755,591 3,267,381 3,209,834 3,685,853

Oct-10 Nov-10 Dec-10

Usage

Supply

Pipeline Econ.

stock

Econ. stock end of week

On-hand end of week

Usage

Supply

Pipeline Econ.

stock

Econ. stock end of week

On-hand end of week


last row indicates the number of products on stock at the manufacturer including the number of

products in transit.

6.4.3 WIP back-end component supplier

The Work-In-Process (WIP) (A3 in Figure 31) is a result of the outcome of the weekly replenishment

calculation. Based on the replenishment calculation the production capacity is balanced over the several

product types. In case the total request is higher than the production capacity in that period a selection

out of the requested plan is made after alignment between the manufacturer and the component

supplier. After alignment the confirmed plan is entered in the model. The supermarket (the customer

decoupling point) is not part of the back-end or calculated in the WIP. One of the parameters for

calculating the safety stock is the lead time. In case the Kanban of the component supplier is not on

target level the lead-time for the back-end will be extend with the lead-time of the front-end. The

second pillar in the safety stock calculation is the uncertainty; in case the supermarket is not meeting

the required standards as calculated in the Kanban system the CV value in the SS calculation should be

revised too.

6.4.4 Generate In-transit overview

The in-transit (A4 in Figure 31) file is an in-between overview for managing the goods flow between

locations. A production batch is in-transit in case a product has left the factory of the component

supplier, but not yet received at the assembly and test manufacturer.

6.4.5 Calculate platelet stock position

Based on the ROP calculation the stock positions at the assembly and test manufacturer are set (A5 in

Figure 31). The stock transactions are administratively managed via an ERP system, the goods received

transaction finishes the in-transit phase (status).

6.5 Results Implementing the new way of working based on the design as written in this master thesis results in a

43.5 M pieces stock reduction (based on 12 M sales per week). The 43.5 M pieces of stock correspond

with 720 K Euro. This enormous stock reduction is a result of two main achievements: lead-time and

variance reduction. With the introduction of the supermarket at the component supplier (customer

decoupling point) and the use of the available supply chain information the reaction time of the

component supplier is reduced from 12 to 1.3 weeks. The variance reduction is mainly caused by

introduction of scheduling rules at the component supplier (first in first out) at the work centers and

buffer lanes. Secondly, one production step at the back-end of the component supplier is outsourced.

Updating the service level agreement with the supplier with logistic parameters like throughput-time,

pick-up /delivery schedules and a day to day communication structure including information sharing

about the midterm (coming 3 months) production schedule reduces supply chain variance. Finally, the

awareness of the effects of being non predictable at the component supplier in terms of stock levels at

the manufacturer brought an focus on doing the right things at the right time according the agreed plan

at the component supplier and the manufacturer as well.


Setting up a control model based on design requirements as set by the organization and the business

characteristics ends with tools to manage the business in a proper way. At the other hand models

without embedded processes at the involved organizations will not bring success after closing the

project (Lee H.L a.o 1992). Figure 34 shows the implemented processes with the designed control model

as backbone and a meeting structure based on terms of references (TOR). The backbone is at the left

side of Figure 34, the monthly and weekly planning sheets are integrated in an overall control process

based on a fix structure. The daily execution is the result of the day to day scheduling based on the

weekly replenishment orders. The backbone starts with a forecast (Sales and Operations Planning),

based on the forecast a production planning is made. This production plan includes also the

performance of the production processes in the period before. Control loops are the linking pins

between the several meetings at monthly /weekly and daily rhythm as well. The performance input is a

bottom up process meanwhile the planning guidelines are top down managed. With the designed

models and the supporting processes supply chain becomes a competitive edge (making the difference).

Figure 34 Planning rules and deploy control model

6.6 Conclusion model design The advice models (ROP, Replenishment, Kanban) are designed to manage specific risks (lead-time and

uncertainties). The models generate advice based on these risks for a desired service level. The desired

service level is expressed in the Z factor. The models give product specific advice for safety stock,

economical stock levels and quantities to produce based on the requested service level and demand.

The outcomes of the models are shown in appendix F, G and H. The outcomes of the model have an

average safety stock of 1.85 week at the assembly and test manufacturer for a service level of 95%. For a

service level of 99% a safety stock of 2.6 weeks is required with a total average stock of 3.1 weeks. The

results of the advice model indicate that a product safety stock level has potential to optimize to the

most suitable stock positions for the Lumileds supply chain at Philips: meaning high customer service

with the right underpinning stock levels in the chain.

Activity

MEETING

MEETING

TOR

TOR

TOR

Day

Week

Month

Production orders

Plan

SCM improvementactions

MonthreportKPI’s

P

D

C

A

WeekreportKPI’s

Daily execution

Review KPI

S&OP ROP settings

DayreportKPI’s

Control ReportPlanForecast

ForecastHistoric sales

Updated

ROP

P

D

C

A

P

D

C

A


Figure 35 shows the ROP development from a financial point of view. The difference between a 90%

service level and a 99.5% service level results in a 23% stock cost level increase with a stable demand

during lead-time and demand and supply variances.

Figure 35 ROP development financially

Finally, with chapter six sub question two (which planning model is most suitable for the LED supply

chain) is answered.

0%

20%

40%

60%

80%

100%

120%

140%

90.0% 92.5% 95.0% 97.5% 99.5%

In K

Eu

ro (

%)

Service level

ROP development

Safety stock

Demand L-T


7 Summary design and modeling phase In this phase a stock level advice model is designed and modeled based on the design objectives and

boundaries as determined in the analysis phase. The Lumileds supply chain is represented as a model

based on three different replenishment strategies; Kanban, Re-Order-Point and Replenishment

calculations. The front-end (all production processes before the customer order decoupling point) of the

component supplier is optimized with the Kanban replenishment strategy. Per product group demand is

translated into Kanban tickets and enabling that the back-end is served with a 99% service level. The

boundaries as set in the Re-Order-Point calculation are leading in the replenishment calculation. Due to

reliable upfront information the inventory required during order lead-time is minimized. The challenge

for the component supplier is to minimize the back-end lead time in such a way that production and

transport lead time is shorter than the throughput time of the parallel process of one of the other key

suppliers of the LED manufacturer.

Designing a most suitable supply chain concept for the component supplier was complex due the fact

that a mass production environment is integrated with a development centre. Production and

development have by nature some conflicting objectives. The supply chain concept must be such that

both are optimally satisfied.

In the analysis phase the design objectives for the model are determined. The realization of these

objectives can be evaluated after the design and modeling phase (Table 1 requirements, criteria,

constraints & design option stakeholders).

With the input of the production planner of the component supplier the supply chain stock

optimization tool is directly linked with the front-end of the component factory. The integration

of the two planning tools enables a minimum of transaction time and is a pro in process

optimization.

One single source for the forecast of the entire supply chain was a key issue for the planner of

the LED manufacturer. With this basic rule all forecast models should use the same basic

information. The lead-time, variance and uncertainties should be worked out as a variable and

are a changeable qualifier in the model.

Design, build and implementation time were important boundaries for the improvement

manager. Secondly the model should be stable and predictable (easy to understand) and finally

calculations of the previous periods should be stored in the model as well.

Simulating different scenarios and reporting were the key topics of the material manager at the

LED manufacturer.

All design requirements and criteria can be evaluated with the model results. The success of the

models is strongly determined by the use in the standing organization. Monthly and weekly

meetings are planned and fully integrated in the business processes of the LED supply chain.


The model designs are based on data of the component supplier and the LED assembly & test

manufacturer. The tools are handed over during three sessions to the end users (the planners at the

component supplier and the manufacturer) and next to this a manual and instructions which will

improve the usability. Possibilities to change /add /remove product characteristics increase the

generality of the tool and give the end user options to make adaptations when necessary. Appendix I

and J show the key information as used in the workshops to introduce the new way of working.

The second phase of design and modeling is finished. In the next and last phase the theoretical models

are transferred into sustainable decision taking models. The last phase ends with conclusions and

recommendations for the future.


Evaluation and validation phase

The design and the models are setup in the analysis, design and modeling phase. The next step is to

evaluate the process and to evaluate in what degree the research has given insights in the research

objectives and questions. In chapter 8 the replenishment strategy models are discussed and

recommendations out of the translation into practice are indicated. The last chapter (9) is reserved for a

validation of the research questions including a personal reflection.


8 From model to practice

8.1 Introduction The designed models are capable to generate a product specific stock level advice including the resulting

replenishment orders or number of Kanban tickets all based on a ROP calculations. The supply chain

uncertainties are covered in the ROP calculation. The outcomes of the models can be used to adapt the

current stock levels to the uncertainties in the supply chain. Better allocation of stock levels can

decrease the total stock level; shifting performance to a new level can improve service. The model

indicates these possibilities theoretically, but to ensure results the model must also be translated into

practice. This chapter describes this translation from model to practice and indicates possible obstacles

and chances. In section 8.2 there is attention to the generality of the model for other situations. Is it

possible to use the models for other situations and which problems can be expected? The production

planner at the manufacturer and at the component supplier both have a very important role in the

translation of the model in practice and must be convinced of the validity and use of the models, as the

conclusions of the stakeholder analysis show. In section 8.3 more about the pilot phase during model

introduction will be discussed. The relation and tension between these stakeholders is part of this

discussion.

8.2 General usability of the model The main goal of this research and the purpose of the model are to generate stock advice, based on lead

time and uncertainties in the supply chain. Combining information about the total supply chain of

Lumileds can help to find an optimal allocation of stocks. With this reallocation a higher service to the

next step in the Supply Chain can be achieved. Besides the main goal, the research and the models are

also a next step in a continuous improvement program to improve ‘customer’ service and lower the

costs. Openness in the Supply Chain is a first step to improve supply chain processes in a bigger context

and is a first step in the direction of collaborative planning. Actually, those models can be used as

support for further cooperation and openness for overall supply chain improvements.

The advice models have a threefold practical purpose:

1. Providing a product specific Kanban advice for the front end of the component supplier.

2. Proving a product specific replenishment stock advice model between the component supplier and

the manufacturer.

3. Supporting tool for further supply chain optimization of the Lumileds supply chain (based on re-

order-point calculation).

Transparency and cooperation between a manufacturer and component supplier have a definite

potential for supply chain optimization. Shifting risks and opportunistic behavior can be a result of the

openness, in the end the supply chain fits in one company, those risks have limited impact at the

bottom-line.


This specific Supply chain optimization issue of Philips Lumileds is, at the end, solved with a standard

design approach based on the theory of Herder and Stikkelman (2004). The models as used within this

assignment are based on proven knowledge and customized to the Lumileds business situation. The

specific elements in this case are the organizational setting and the interdependencies between the

assembly and test manufacturer with the key component suppliers. In this situation the key component

suppliers are all part of one company. Due to the intercompany relationship between the suppliers and

the manufacturer the way of managing a sub contractor wasn’t strictly (according agreed guidelines

saved in a contract) arranged in execution from the past. Creating meeting rules (TOR, RACI) and setting

up decision criteria when calculating the ROP or replenishment orders were part of the maturity shift of

the Lumileds organization. The maturity shift was a need to implement a structured way of working

based on companywide agreements.

The applicability of the design model in industries outside Lumileds with or without seasonality is

enormous due the fact that the models are running on two planning cycles, the monthly and the weekly

planning cycle. The monthly planning cycle brings an outlook for the coming four quarters and based on

the MAT (moving average total) the demand as used in the next calculation steps takes already future

development into account. Based on the calculated system boundaries in the monthly planning process

the weekly replenishment orders are calculated between a component supplier and the manufacturer.

The setup with the monthly and the weekly planning cycles makes the design model as a generic

applicable supply chain control model for the product creation industry.

8.3 Using the model for a pilot

The advice models are based on theory and input from the component supplier and the LED assembly

and test manufacturer. Those models and the research where the models are based on, has provided

insights in the supply chain uncertainties. To improve the suitability for Lumileds the model is extended

into a tool. This tool fits well with the current supply chain processes of Lumileds and makes is possible

to adapt product specific characteristics.

Product and characteristics can change over time and maintenance is necessary to keep the tool useful.

The tools are handed over to the production planners of the component supplier and the LED

manufacturer. During workshops the handover was formalized in combination with a manual on how to

use and maintain the tools in the future. The manual describes the design of and theory behind the

replenishment strategy models and also explains how to make changes.

During the pilot phase (1 month) the models were used with 2 product families. Based on the

experiences of the field experts modifications were made before we extend the pilot with another

month. During the second pilot 4 extra product families were included. Based on the results out of the

second pilot the models were finalized and handover to operations ready. The last part of the hand-over

to operations was to imbed the models in the standard organization. Term-Of-References (TOR) were

made for creating meeting guidelines and meeting inputs & outputs /results. With the help of the TOR a

structure was set for having efficient and fruitful weekly and monthly alignment meetings. As a direct


effect of the transparent meeting structure thinking in improvement opportunities were embed in the

day to day business easily.

Effects of the pilot on the stock positions were directly visible. Due to the information sharing and the

reduced lead-time no production was necessary from the component supplier in the first week after the

changeover (start of the pilot). After the first week the replenishment values were positive, meaning

new components were needed to fill the stock positions to the requested levels. After each introduction

or update of the ROP calculation the stock positions were discussed with the stakeholders during the

ROP meeting, actions were taken in the replenishment meeting. With this setup the content discussions

were based on boundaries as set during the ROP meeting. The ROP meeting was full of discussions

about demand and supply variances, batches sizes (yield), throughput-time and the total demand plan

for the coming period.

The design model of Herder and Stikkelman (2004) has a focus on design and less at the implementation

part of a project. The assignment of this master thesis was to design and to implement supply chain

controls. For that reasons the design model of Herder & Stikkelman (2004) was enriched with the

approaches of Shannon (1998) and Lee & Billington (1992) to enable the project to use an integrated

approach from the start of the project till implementation and finally a sound handover to operations.

The scope of the approach of Herderand Stikkelman (2004) is unlimited: additional building blocks can

be added (f.e. number of iterations) when desired. The flexibility of the model and on the other hand

the step by step approach was useful during the design and development phase of the projects. The

focus on the design parameters and the actor analysis during and before the design phase enables a

smooth implementation. The lack of structure in the implementation part of a model is the weak point

of the approach of Herder and Stikkelman (2004). No structure or implementation guidelines are

provided in the approach. For that reason additional research was necessary for closing the gap and to

run the project on a proven knowledge based design and implementation approach. For next design and

implementation projects the opportunity is there to enrich and integrate the design model of Herder

and Stikkelman (2004) with a standard implementation approach of (change) processes. The integration

of a design and implementation model was out of scope of this research /master thesis.

Table 1 shows the requirements of the most important stakeholders. Based on the stakeholders input

and the characteristics (no seasonality, fast growing and demanding customers) of the Lumileds

business the design phase was started. Evaluating the model characteristics based on the requirements

of the stakeholders (as provided by table 1), the implemented supply chain control model and the

related business control processes are from all perspectives better than initial requested. Strategic

behavior of one of the key stakeholders can always change the input of table 1, but will not harm the

designed and implemented control model because the business characteristics of the LED industry were

the first attention points; secondly the requirements of the stakeholders were checked on completeness

necessary for acceptation and hand-over to operations. In case a key stakeholder will change their

mindset it is not that difficult to change some output parameters, in this: changing a mindset will not

change a business setup.


8.4 Summary from model to practice In this chapter the translation from the model into practice is described. The designs of the models are

based on theory and implemented with the field experts within the Lumileds supply chain. The designed

models are turned into tools for Philips Lumileds. Besides the several hand-over sessions a tutorial is

written were all possibilities to change /added /delete product are described.

Within this chapter sub question 6 (What actions are recommended to convince the actors about the

benefits of the new supply chain planning model and way of working) is answered. Using a pilot for a

limited part of the business was the right way to do in convincing stakeholders. With the results of the

pilot it was easy to share the benefits of an optimal supply chain control model.


9 Conclusion and recommendations

9.1 Introduction In the previous chapter the model design is evaluated and the model results are validated. In this

chapter conclusions are drawn up based on the results of this research. The research questions

determined the direction and structure of the research. In the conclusions answers are provided on the

research questions. The research questions therefore form the structure of the conclusions section (9.2).

Recommendations are provided in section 9.3 for aspects out of the research scope and aspects which

deserve further research.

9.2 Conclusions of the research Supply chain optimization based on customer service and cost minimization is the most important

drivers for Philips Lumileds. Improving the service will result in fewer out-of-stocks at the back-end of

the component supplier and the manufacturer. This research has one objective.

Generate, implement and deploy supply chain planning controls for shop floor scheduling and

material replenishment for the LED products between a component supplier and the LED assembly and

test manufacturer.

The stock level optimization results in better service in the supply chain, which can be measured in cycle

time, variance and availability. The ROP model can be used for realizing service improvement by

managing the stock levels accordingly. The performed research gives an answer on the main research

question:


SAFETY STOCK LEVELS BETWEEN THE COMPONENT SUPPLIER AND THE LED ASSEMBLY & TEST MANUFACTURER.

Before an answer on the main question can be formulated, first the sub-questions will be answered,

based on the performed research. The sub-questions as set in section 1.2.3 will now be discussed and

answered systematically.

1. Which facts determine the most suitable supply chain planning method?

In the current situation the LED assembly and test manufacturer send in orders every week. The

order frequency is based on the demand pattern of the industrial customers of the manufacturer.

The difference per product makes that the models are generating a separate and specific advice per

product per part of the Supply chain as indicated in Figure 17 (section 5.1). The safety stock levels

are not only determined by uncertainties but also by the desired service level. In this research the

models generates advice for a service level of 97.5% (default) but can be manual changed. Searching

for the most suitable planning model is for the total Lumileds Supply chain ends in a model

configuration of best practices per part of the chain. A Kanban replenishment strategy system for

high volume and low costs products, replenishment orders based on supply chain stock availability


to supply the manufacturer and an overall control based on Re-Order-Point calculation. Main inputs

for the three models are the Sales & Operations Planning, supply chain lead-times and the various

supply chain uncertainties.

2. Which planning model is most suitable for the Lumileds supply chain?

A most suitable planning method is determined by several aspects. Managing stock levels and Supply

Chain uncertainties are two aspects to cover. The Lumileds supply chain can be grouped into three sub

models, which are all influencing the stock positions at the LED assembly & test manufacturer (section

4.3).

Figure 36 Concept advice model

These three sub models in combination with a desired service level in the Supply Chain determine the

stock levels at the LED assembly & test manufacturer. The front-end (the processes at the component

supplier before the decoupling point) is controlled by the forecasted demand pattern translated into a

Kanban calculation at article level. Uncertainties (demand, lead-time and batch size) in this part of the

model are translated into factors in the formula. In the Re-Order-Point calculation the boundaries are

set for the overall component flow, meaning; replenishment levels with respect to the work in process

at the component supplier after the decoupling point and the in-transit. Uncertainties covered in the

ROP calculation are; demand variance, supply lead-time variance and supply quantity variance. In the

weekly replenishment process the output of calculations are used as set in the ROP. The supermarket

between the front- and back-end (decoupling-point) guarantees a 99% availability of available products

in relation with the replenishment orders. Based on literature and experts a list is created of

characteristics which possibly influence the choice of a supply chain control method.

The major characteristics which determine the planning method are:

Supply chain lead-time

Volume

Responsiveness component supplier

Demand variation

Supply variation

Re-order-point

calculation

Replenishment

orders

Stock

levels

Manuf

acturer+/- +/-

S&OP

Supply variance

Demand variance

Moving average

Lead-time

Service level

Product life-cycle

phase

Stock take

WIP

In-Transit

ROP

Kanban

component

supplier

S&OP

Supply variance

Lead-time

Batch size

Service level

+/-

+/-


Stock value

3. Which processes are related to the supply chain planning processes and what are the

interdependencies?

To generate product specific advice the overall planning cycle should be incorporated. Figure 37 shows

the relationship between the two main component suppliers of the LED assembly and test manufacturer

which shows a (partly) parallel planning opportunity. Making component supplier1 leading in the Supply

chain saves 10 production days to be covered by stock including the uncertainties involved.

Week 1 Week 2 Week 3 Week 4

Assembly & test

manufacturer



Assembly & test

PP

PP

Production &

transport

Production & transport

PP = production plan

Difference in production lead

time to cover with stock

Figure 37 Planning of replenishment orders

Secondly, the ROP calculation affects the two other planning tools directly with a weighted average

demand for the coming 2 till 3 months and replenishment levels besides the overall used CV values for

covering the Supply chain uncertainties.

4. What are the relevant actors, what role do they currently have and what role should they have

according the new model?

In the current situation the production planner of the component supplier is managing the stock levels

of the LED assembly and test manufacturer. Due to the implementations of the Supply chain control

models the planning method is changed into a format in which the stock owner (production planner of

the manufacturer) is controlling the Supply chain from the decoupling point of the component supplier

till the goods are on stock at the LED assembly & test manufacturer.

Roles and responsibilities are changed in this new control setup (see also appendix G) from a known

uncontrolled situation to a known controlled situation. In this new setup the production planner of the

component supplier and the LED assembly & test manufacturer are the first in line for managing the day

to day operations. Escalation in case an issue is not solvable within the boundaries which are set in the

ROP is in line with the organization hierarchy.


Figure 38 stakeholder diagram

Figure 38 shows the overview of the involved stakeholders. All critical marked stakeholders have any

relationship with the new control setup according the RACI overview (as presented in appendix G).

5. Which key performance indicators are relevant to evaluate the supply chain plan model and

what are the critical success factors?

Based on research and the outcome of discussions with field experts the following metrics are

introduced:

1. CT measurement per sub process including the variation and target settings

2. Stock levels (at the component supplier (decoupling point) and at the LED assembly and test

manufacturer) are judged against the safety stock levels.

3. Delivery performance component supplier towards the LED assembly & test manufacturer at

component and quantity level.

Above metrics are new in the supply chain of Lumileds and are monitored at a weekly basis. Actions are

taken out of the measurement and filtered on incidents and structural (repeating) failures. A key

element in those measurements is besides a reliable data source and ownership continuous

improvement cycles in which the process improvements regarding cycle time and variance reduction are

part off.

6. What actions are recommended to convince the actors about the benefits of the new supply

chain planning model and way of working?

Benefits for all related actors are the best arguments to convince stakeholders about the improvement

possibilities in there Supply chain. Nevertheless, besides selling all the goods out of the advice models

different workshops were organized to inform all stakeholders about the supposed changes before we

CriticalNon-Critical

Dedicated

Non-Dedicated

Production planner

Maarheeze

Production

manager

Maarheeze

Engineering

manager

Maarheeze

Development

manager

Maarheeze

Product manager

MaarheezeSupply manager

PenangProduction planner

Penang

Component planner

Penang

Production

manager Penang

Supply chain

excellence

manager

Planning manager

Penang


went live with the new planning concept. The results out of the pilot were useful examples to explain

the new ideas and to show the actors the effects based on recognized system information. After

introducing of the new concept an after care (managed by experienced consultants) period of eight

weeks was arranged upfront to safeguard the organization for any fall back due to lack of knowledge

regarding the tools and /or calculations rules.

With help of the sub-questions it’s possible to formulate an answer on the main question:


SAFETY STOCK LEVELS BETWEEN A COMPONENT SUPPLIER AND THE LED ASSEMBLY & TEST MANUFACTURER.

The advice models gives a product specific advice based on the different lead-times and the Supply chain

uncertainties including a specific demand pattern. Based on the current situation the models are

calculating an advice related to the front-end of the component supplier. The back-end of the

component supplier is controlled via a replenishment calculation al managed by a Re-Order-Point

calculation. The ROP and Kanban boundaries are set ones a month, the replenishment orders are

calculated weekly. As a result of above structure the organization is able to measure the different parts

of the Supply chain and the overall Supply chain as well. The measurements about the sub parts of the

Supply chain are a result of the new control model. The models results in transparency and the

opportunity to simulate the impact of variance or lead-time reduction on stock levels. Based on the

simulation results the model support improvement potential and proves that an equal service is possible

with lower stock levels. The influence of the component supplier and the LED assembly & test

manufacturer can be used to improve and focus on cycle time reduction and controlling the

uncertainties in the Supply chain as much as possible. Finally, the models are the first step to further

improvements & cooperation between the component supplier and the manufacturer.

9.3 Recommendation for further research Although this research project is finished; the insights of this research have identified other future work

which can be performed. These recommendations are divided into three recommendations for future

research.

The first and most important recommendation is to establish, improve and extend the relation between

the component supplier and the LED assembly & test manufacturer. As described in the previous

chapter the advice models are one step in the right direction of a fully controlled Supply chain. The next

step is steering together on improving performance indicators and to start up a variance reduction

program. After that Supply chain integration can be one of the next steps to go in further optimization

and drive together integral Supply chain improvement projects. In a couple of years from now the

situation can be ready for the next steps into system integration between the component supplier and

the manufacturer. Working in one system environment will bring a closer cooperation and opens

especially options for examining planning and forecast root causes like demand /supply

underestimation, phase in or phase out of products and elephant orders for special customers.


The second recommendation is aimed at the goods supply at the component supplier. The source part

of the component supplier was not in scope of this research. It would be worthwhile to also investigate

in the order pattern and purchasing activities in the several articles which are used at the component

supplier. Using the Philips buyer power and setting up system control limits will enable the component

supplier to enter the next level Supply chain control at their source side.

The last recommendation is an extension of the models with a sensitivity check on the several CV factors

and lead-times with the effects on the safety stock levels. Stock and service level decisions are related

with cost considerations. Costs are not part of the model yet. The models can support decision taking

based on service and Supply chain uncertainties, implementing a cost aspect improves the quality of the

decision support power of the models.

9.4 Reflection

9.4.1 Reflection on theory and methodology

A benefit of the design set up is that the models including the way of thinking behind can be indicated as

a general approach to solve comparable issues. The general approach based on the model of Herder and

Stikkelmans (Figure 4) methodology was a guideline to organize researches like this. A missing part of

the methodology of Herder and Stikkelmans is the implementation approach. Therefore input of the

field experts and the publication of D.J. Bowersox ao, 1999 – managing introduction risk through

response-based logistics was added to shape the implementation part of the project.

9.4.2 Reflection on results

Creating a tool for optimizing the controls in the Supply chain was the assignment. Finally a theory based

model was designed and implemented in the Lumileds supply chain. After two iterations the final

version of the models was implemented within the timeframe of the aftercare period. In the last four

months the models are running quite smoothly and are embedded in the running organization. The

models have drastically improved business decision taking. The results of this project move the Lumileds

Supply chain control to a higher platform, by managing the short- and medium term supply and demand

balance.

9.4.3 Personal reflection

The understanding and knowledge gained in my personal development during this research was

enormous. Working with knowledgeable people during this research was really fun and brought me

besides this Master Thesis to a new direction of using science articles for solving day to day business

issues.


References Affonso R. , Grabor B., Marcotte F.(2009). Coordination Model for Supply Chain Management, University

of Toulouse.

Baganha M. P. and Cohen M. A. (1998). The Stabilizing Effect of Inventory in Supply Chains, Operations

Research, Vol. 46, No. 3, Supplement: Focus Issue on Manufacturing Logistics (May - Jun., 1998), pp.

S72-S83.

Bowersox, D.J. (1990). The strategic Benefits of Logitics Alliances. Harvard Business

Review, 86 (4):p.36-43.

Bowersox D.J., Stank T.P. and Daugherty P.J. (1999). Managing introduction risk through response-based

logistics.

Chopra, S., Dada, M. and Reinhardt, G. (2004). The effect of lead time uncertainty on

safety stocks. Decision sciences, 35 (1):p.1-24.

Corsten, D. and Gruen, T.W. (2004). Stock-Outs Cause Walkouts. Harvard Business

Review:p.26-28.

Davis T. (1993). Effective supply chain management.

Enserink, B., Koppenjan, J.F.M., Thissen, W.A.H., Kamps, D.P. and Bekebrede, G.

(2008). Analyse van complexe omgevingen, Delft: Delft University Press.

Eppen, G.D. and Martin, R.K. (1988). Determining safety stock in the presence of

stochastic lead time and demand. Management Science, 34 (11):p.1380-1390.

Fisher, M.L., Hammond, J.H., Obermeyer, W.R. and Raman, A. (1994). Making Supply

Meet Demand in An Uncertain World. Harvard Business Review, 72 (3):p.83-93.

Fleisch, E. and Tellkamp, C. (2005). Inventory inaccuracy and supply chain

performance: a simulation study of a retail supply chain. International Journal of Production

Economics, 95 (3):p.273-385.

Graham C. Stevens (2007). Integrating the Supply Chain, Peat Marwick McLintock, London.

Herder, P.M. and Stikkelman, R.M. (2004). Methanol-based industrial cluster design: a

study of design options and the design process. Industrial & Engineering Chemistry Research,

42:p.3879-3885.


Hengst M. den and Herder P.M. (2004). Quick and Dirty’ Modeling in a Decision Support Tool for Supply

Chain Design, international Confcrence on Systems, Man and Cybernetics 445-450.

Jammernegg, W. and Reiner, G. (2007). Performance improvement of supply chain

processes by coordinated inventory and capacity management. International Journal

Production Economics, 108:p.183-190.

Johson J. D. (2005). Managing variability in the semiconductor supply chain.

Kok T.G. de and Fransoo J.C. (2002). Planning supply chain operations: definition and comparison of

planning concepts.

Kok T. de, Janssen F., J. van Dorenmalen, E. van Wachem, M. Clerkx and W. Peeters (2005). Philips

Electronics Synchronizes Its Supply Chain to End the Bullwhip Effect.

Lee H.L. and Billington C. (1992). Managing supply chain inventory: pitfalls and opportunities.

Lutza S., Loeddingb H., Wiendahl H-P. (2003), Logistics-oriented inventory analysis, Int. J. Production

Economics 85 (2003) 217–231.

Marmostein, H. and Zinn, W. (1990). Comparing two alternative methods of

determining safety stock levels: the demand and the forecast systems. Journal of Business

Logistics, 11 (1):p.95-110.

H. Mintzberg (1983). Structure in Fives: designing effective organizations: p. 121-156.

Orcun S., Çetinkaya S. and Uzsoy R. (2007). Determining safety stocks in the presence of workload-

dependent lead-times.

Ortmanns J., Mooren J., Rulkens H., Mingels J. (2010). Hikari industrialization assesment.

Shannon R. E. (1998). Introduction to the art and science of simulation, Industrial Engineering, Texas

A&M University, College Station, Texas 77843-3131.

Schneider, M. (2009). Out-of-Stock out of business, Enschede, TU Twente.

Sanajian N. and Balcıoglu B. (2009). The impact of production time variability on make-to-stock queue

performance, European Journal of Operational Research 194 (2009) 847–855.

SUWANRUJI P. and ENNS S.T. (2005). Evaluating the effects of capacity constraints and demand patterns

on supply chain replenishment strategies, International Journal of Production Research, Vol. 44, No. 21,

1 November 2006, 4607–4629.


Tien-Yu Lin (2010). An economic order quantity with imperfect quality and quantity discounts. Applied

Mathematical Modelling 34 (2010) 3158–3165.

Zhao, X., Xie, J. and Leung, J. (2002). The impact of information sharing and orderingco-ordination on

supply chain performance. Supply Chain Management: an International

Journal, 7 (1):p.24-40.

Zinn W. and Marmorstein H. (1990). Comparing two alternative methods of determining safety stocks

levels: the demand and the forecast systems.


Appendices

Supply chain planning at Philips Lighting Lumileds

How secure do we like to be?


end 2 end environment to improve product availability.

Author:

R. Hartevelt

Confidential


Appendix A Product families

Product family SKU component

supplier

SKU assembly & test

manufacturer

Luxeon Flash PWF5 3222 023 62010 LL60.0163 Bin 330

3222 023 63010 LL60.0163 Bin 440

Luxeon Flash PWF6 3222 023 63310 LL60.0134 Bin 222

3222 023 63410 LL60.0134 Bin 333

3222 023 63510 LL60.0134 Bin 555

3222 023 63610 LL60.0134 Bin 444

Luxeon Rebel Warm White 3222 023 61200 LL60.0187 Bin 231

3222 023 61200 LL60.0187 Bin 232

3222 023 61200 LL60.0187 Bin 241

3222 023 61200 LL60.0187 Bin 242

3222 023 61200 LL60.0187 Bin 251

3222 023 61200 LL60.0187 Bin 252

3222 023 61200 LL60.0187 Bin 261

3222 023 61200 LL60.0187 Bin 262

Luxeon Altilon C2 3222 023 60100 LL60-156-018 Bin 1820

3222 023 60400 LL60-156-018 Bin 1813

3222 023 60300 LL60-156-018 Bin 1806

Luxeon Rebel PC Amber 3222 023 64200 LL60.125 Bin 3 - 5

Luxeon Rebel Hikari 2700K 90CRI T&R 3222 023 6500 LL60.0171 Bin 3 – 5

Luxeon Rebel Hikari 3000K 80CRI T&R 3222 023 65100 LL60.0174 Bin 3 – 5

Luxeon Rebel Hikari 4000K 70CRI T&R 3222 023 65200 LL60.0177 Bin 3 - 5


Appendix B Relation between stock level and service level

Correlations

Stock level Customer line item performance

Stock level Pearson Correlation

N

1,000

26

0,072

26

Customer line item

performance

Pearson Correlation

N

0,072

26

1,000

26

Table 4 Results correlation analysis stock level and CLIP

The correlation between stock level and CLIP can help to control if a higher stock level results in a better

CLIP. There is a significant and relation between stock level and CLIP. The relation is 0.072, which means

a high correlation coefficient.


Appendix C Philips analysis Royal Philips Electronics of the Netherlands is a diversified Health and Well-being company, focused on

improving people’s lives through timely innovations. As a world leader in healthcare, lifestyle and

lighting, Philips integrates technologies and design into people-centric solutions, based on fundamental

customer insights and the brand promise of “sense and simplicity”. Headquartered in the Netherlands,

Philips employs approximately 116,000 employees in more than 60 countries worldwide. With sales of

EUR 23 billion in 2009, the company is a market leader in cardiac care, acute care and home healthcare,

energy efficient lighting solutions and new lighting applications, as well as lifestyle products for personal

well-being and pleasure with strong leadership positions in flat TV, male shaving and grooming, portable

entertainment and oral healthcare.

Sustainability is at the center of Philips’ strategy. Philips is committed to reducing its environmental

footprint in all aspects of its business: in the products, manufacturing, procurement, as well as in the

communities where the company acts and in the working practices of its employees. All Philips products

go through an EcoDesign process, identifying environmental impact in terms of energy efficiency,

hazardous substances, take-back and recycling, weight and lifetime reliability. Philips’ processes on

Green Product sales are verified annually by an independent third party and published in the Annual

Report. Philips aims to combat global healthcare challenges by focusing on delivering better quality

healthcare at lower costs, also in the emerging markets, such as China and India. Philips also takes a

leading position in educational programs, showing its stakeholders that energy efficient solutions are

simple, easy and actionable and make economic sense for national and local governments, businesses,

schools and individuals.

Philips has 3 sectors (Healthcare, Lighting and Consumer Lifestyle) a design- and a research unit. In 2009

Philips invested EUR 1.6 billion in research and development.

Philips Lighting

Philips Lighting is a leading provider of solutions and applications for both professional and consumer

markets. We address lighting needs in a full range of environments – indoors (homes, shops, offices,

schools, hotels, factories, and hospitals) as well as outdoors (public places, residential areas and sports

arenas). We also meet people’s needs on the road, by providing safe lighting in traffic (car lighting and

street lighting). In addition, we deliver light-inspired experiences through architectural and city

beautification projects. Our lighting is also used for specific applications, including horticulture,

refrigeration lighting and signage, as well as heating, air and water purification, and healthcare. With the

new lighting technologies, such as LED technology, and the increasing demand for energy efficient

solutions, Philips will continue shaping the future with groundbreaking new lighting applications.

Philips Lighting Lumileds

Philips Lumileds Lighting Company (founded 1999) is the world's leading manufacturer of high-power

LEDs and a pioneer in the use of solid-state lighting solutions for everyday purposes including

automotive lighting, computer displays, LCD televisions, signage and signaling and general lighting. The

company's patented LUXEON® Power Light Sources are the first to combine the brightness of


conventional lighting with the small footprint, long life and other advantages of LEDs. The company also

supplies core LED material and LED packaging, manufacturing billions of LEDs annually, and ranks as the

producer of the world's brightest red, amber, blue, green and white LEDs.

Lumileds began as the optoelectronics division in Hewlett-Packard (HP) almost 40 years ago. Hewlett-

Packards' experts literally wrote the book on LEDs. In the late 1990's, recognizing the potential for solid-

state lighting, HP and Philips, one of the world's leading lighting companies, began exploring how they

could work together and deliver a new solid-state lighting solution to the market. In 1999 HP split its

company into two, and the optoelectronics group was assigned to the new Agilent Technologies. In

November of the same year, recognizing the enormous potential for LEDs, Agilent Technologies and

Philips formed Lumileds and assigned it the responsibility of developing and marketing the world's

brightest LEDs and enabling a new world of light. In 2005, Philips acquired Agilent Technologies’ interest

in Lumileds. Today, the company continues to lead the industry in the development and release of

increasingly brighter and technically advanced solid-state lighting technology.


Appendix D Details sub processes front-end component supplier

Monthly update based on

customer demand Penang

KANBAN

MES replenishment report

SlurryFIFO

GranulationFIFO

UAP BBO SinterFIFO FIFO

Pre-GrindingFIFO FIFO

WIPOverview_replenishmentorders

Median = 19:12

Min = 8:48

Max = 62:42

Buffer-time = 0

Yield = 100%

OXOX

Median = 17:18

Min = 2:36

Max = 63:30

Buffer-time = 17:42

Yield = 88.13%

Median = 1:18

Min = 0:30

Max = 2:12

Buffer-time = 166:00

Yield = 99.83%

Median = 25:46

Min = 20:46

Max = 45:36

Buffer-time = 36:54

Yield = 99.95%

Median = 55:24

Min = 37:36

Max = 83:06

Buffer-time = 0:12

Yield = 99.95%

Median = 4:00

Min = 2:30

Max = 12:48

Buffer-time = 23:30

Yield = 99.18%

PWF6 & HIKARI

Process Time: 122:40 hours

VSM Lead Time: 244:54 hours

19:12 hr 17:18 hr 1:18 hr 25:46 hr 55:24 hr 4:00 hr


Appendix E Details sub processes front-end component supplier

Weekly update based on planned

consumption & stock take Penang,

in-transit and Re-Order-Point

calculation

Replenishment

GrintapingFIFO

GrindingFIFO

Measuring DCF Coat & CureFIFO FIFO

SepTap

Median = 2:06

Min = 0:42

Max = 4:12

Buffer-time = 0

Yield = 99.82%

Median = 15:42

Min = 2:48

Max = 24:48

Buffer-time =13

Yield = 93.72%

Median = 0:12

Min = 0:06

Max = 5:54

Buffer-time = 0

Yield = 100%

Median = 120

Min =

Max =

Buffer-time =

Yield = 90%

Median = 33:24

Min = 4:54

Max = 110:54

Buffer-time =

Yield =

Median = 3:42

Min = 1:36

Max = 11:00

Buffer-time = 3:42

Yield = 81.56%

HIKARI



Seperation Visual Insp. CTISFIFO FIFO FIFO FIFO

Median = 10:12

Min = 2:12

Max = 16:36

Buffer-time = 4:18

Yield = 100%

Median = 10:24

Min = 3:48

Max = 34:48

Buffer-time = 41:00

Yield = 100%

Median = 35:54

Min = 25:24

Max = 50:00

Buffer-time = 6:12

Yield = 100%

OXOX

2:06 hr 15:42 hr 0:12 hr 120 hr 33:24 hr 3:42 hr 35:54 hr 10:24 hr 10:12 hr

GrintapingFIFO

GrindingFIFO

Measuring SepTap

Median = 1:42

Min = 0:12

Max = 4:12

Buffer-time = 0

Yield = 99.73%

Median = 13:06

Min = 1:42

Max = 30:18

Buffer-time = 10:00

Yield = 80.72%

Median = 0:06

Min = 0:06

Max = 24:48

Buffer-time = 0

Yield = 100%

Median = 1:12

Min = 0:30

Max = 4:24

Buffer-time = 26:36

Yield = 77.84%

PWF6



Seperation Visual Insp. CTISFIFO FIFO FIFO FIFO

Median = 6:30

Min = 0:24

Max = 30:18

Buffer-time = 2:30

Yield = 100%

Median = 3:48

Min = 0:30

Max = 33:30

Buffer-time = 7:30

Yield = 100%

Median = 32:06

Min = 7:12

Max = 145:24

Buffer-time = 5:00

Yield = 100%

Weekly update based on planned

consumption & stock take Penang,

in-transit and Re-Order-Point

calculation

Replenishment

OXOX

1:42 hr 33:24 hr 0:06 hr 1:12 hr 32:06 hr 3:48 hr 6:30 hr


Appendix F Model results Kanban front-end component supplier

Product family Product Demand CV

Lead-time

front-end Batch size

Lead-

time

demand

Safety

stock

95%

Safety

stock

95%

(days)

Total

Kanban

95%

Total

Kanban

95%

(days)

Safety

stock

97.5%

Safety

stock

97.5%

(days)

Total

Kanban

97.5%

Total

Kanban

97.5%

(days)

Safety

stock

99%

Safety

stock

99%

(days)

Total

Kanban

99%

Total

Kanban

99%

(days)

Luxeon Flash5 3222 023 62010 228,538 0.34 2.94 110,000 7 2 9 2 9 2 9

Luxeon Flash5 3222 023 63010 533,256 0.34 2.53 110,000 13 3 16 4 17 4 17

Luxeon Flash6 3222 023 63310 116,751 0.71 2.34 117,000 3 2 5 2 5 2 5

Luxeon Flash6 3222 023 63410 817,259 0.71 2.10 117,000 15 9 24 10 25 12 27

Luxeon Flash6 3222 023 63610 1,050,762 0.71 2.14 117,000 20 11 31 13 33 15 35

Luxeon Flash6 3222 023 63510 350,254 0.71 2.17 117,000 7 4 11 5 12 5 12

Luxeon Rebel WW 3222 023 61200 98,129 1.36 3.43 90,000 4 3 7 3 7 4 8

Luxeon Rebel Hikari 2700K 80CRI T&R 3222 023 65000 91,847 1.00 3.47 60,000 6 3 9 4 10 4 10

Luxeon Rebel Hikari 3000K 80CRI T&R 3222 023 65100 147,404 1.00 3.56 60,000 9 5 14 5 14 6 15

Luxeon Altilon C2 3222 023 60100 11,886 0.79 2.93 40,000 1 1 2 1 2 1 2

Luxeon Altilon C2 3222 023 60400 62,401 0.79 2.61 40,000 5 3 8 3 8 3 8

Luxeon Altilon C2 3222 023 60300 72,801 0.79 2.36 40,000 5 3 8 3 8 4 9

Luxeon Rebel PC Amber 3222 023 64200 159,863 0.84 6.00 95,000 11 3 14 3 14 4 15

3,741,151 106 52 5.2 158 15.8 58 5.8 164 16.4 66 6.6 172 17.2


Appendix G RACI model Supply chain control model

RACI ROP process lumiramics

Activity /Decision in process

SC im

prove

men

t man

ager P

enang

Mate

rial m

anage

r Penan

g

Inte

gral p

lanner M

aarheeze

Plate

let p

lanner

Penan

g

Operatio

ns man

ager l

umira

mics

Mhz

Sr. e

ngineer

ing p

lannin

g offi

cer P

enang

WW

supply

pla

nning

man

ager

Monthly process1 Orientation sheet

Update S&OP R/A C I I

Update bin distribution C A I R

2.1 Copy JDE Sales order detail report into Monthly ROP A R

2.2 Verify Shipped table on completeness A R

2.3 Shipped Pivot

Update product family A R

Determine selection period A R

3 Determine average demand period A R

4 ROP calculation

Determine service level C R/A I I

Determine average lead time back-end Mhz I R/A I

Determine average lead time variation back-end Mhz I R/A I

5 Update Product Life Cycle overview A I I R

Weekly process0 Prepare KPI reports

Penang part A R

Maarheeze part R A

1a Copy JDE reports Stock Aging report into weekly ROP A R

1b Verify Platelet stock table on completeness A R

1 Copy Stock take values into table A R

2 Determine WIP back-end Mhz R A

3 In transit

Update shipped goods R A

Copy JDE report received goods A R

4 Copy ROP from monthly to weekly ROP sheet A R

5 MD update A R

6 Copy BP Penang into weekly ROP A R

7 Copy # of GGI starts into weekly ROP A R

8 Verify GGI history sheet on completeness A R

9a Verify Calc replenishment on completeness /rightness A R

9b Verify Production Orders MHZ on completeness /rightness A R

General activities1 Maintenance of weekly & monthly ROP calculation sheets A C C C I R

2 S&OP forecast Png -> Mhz R/A I I I I

3 Capacity commitment process Mhz -> Png I I R I A

4 Escalation process capacity commitment A C R R A

Allocation decisions to deal with short term constraints R C C C C A

Plan to increase capacity to fulfil capacity need on mid term

Inventory build up decission to avoid capacity constraints

Phase in /phase out products in weekly /monthly process C A C R I I

5 yearly process review /maintenance platelet planning A R C C I C

Legend:

Legend:

- Improvement & Process Responsibility - Process/Improvement Consult

- Process Accountability - Process/Improvement Inform

R

A

C

I


Appendix H IDEF0 level 2

TITLE:NODE: NO.: 2level A1 SADT Lumileds /lumiramics

S&OP

Finished product

shipments

Service level

MAT

Lead-time platelets

Product-life-

cycle

A1.1

Transform

orientation from

monthly to

weekly

A1.2

Shipped data

consolidated per

prod family

A1.3

Calculate

average platelet

demand

A1.4

Calculate

demand variation

A1.5

ROP calculationRe-order-point

Selection period


Weekly bin

forecast

Platelet

demand

Weekly qty

per prod

family

Demand variation per prod

family

Prod family table


WIP MaarheezeIn-transit

Consolidated stock overview

GGI start

ROP

A2.1

Replenishment

calc.

Confirmed plan

Yield back-end Maarheeze

A2.2

Production

planning MHZ

Cap

constraints

Production request



Stock pre grinded wafers

Confirmed plan

A3.1

Planned orders

MES

Goods receipt production order

Supermarket front-end

Goods Issue back-end Mhz

Production

planning

A3.2

Back-end

production Mhz

Confirmed plan


Carrier lead-time

Platelets

A4.1

Pack goods

A4.2

Transport

A4.3

Prepare platelet

stock report PngStock data overview

A0

Generate In-

transit report

Platelets on transport

Platelets received

Platelets In-transit

Goods

Ready

for transport



Platelets In-transit

JD Edwards stock data

ROP

Aggregated stock data overview

A5.1

generate platelet

stock report


Appendix I Monthly Re-Order-Point process

Confidential Philips Applied Technologies – Industry Consulting, Juli 2010

Monthly process

1. Review product life-

cycle

2. Prepare orientation

sheet

3. Download shipped data

from JDE

4. Determine selection period CV

5. Determine moving average period

6. Update average LT

& LT var and determine

service level

7. Copy new ROP values

in weekly ROP sheet

8


Step 1a. Review product life-cycle

• Output of S&OP meeting (tab 5.)

9

1. Review product

life-cycle2.

Prepare orientation sheet

3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update

average LT & LT var and

determine

service level

7. Copy new ROP

values in

weekly ROP sheet


Step 1b. Product family determination

• At tab 5

– All used item numbers are listed including the product family name

– Like:

– Based on the translation table the product families are grouped in

the pivot

10

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet

Item Number Product Family

LXCL-PWF4N Flash PWF5

LXCL-PWF4W Flash PWF5

LXCL-PWF4W-3000 Flash PWF5

LXCL-PWF5 Flash PWF5

LXCL-PWF5D Flash PWF5

Sum of Quantity Shipped Column Labels

Row Labels 1008 1009

Altilon 4,266 10,038

Flash PWF5 250,000 310,760

Flash PWF6 2,800 1,010,200

Rebel PC Amber 310 135,090

Rebel WW Lumiramic 11,500 58,000

Platelet, Hikari 2700K 80CRI


Grand Total 268,876 1,524,088



Step 2. Prepare orientation sheet

• Lumiramics forecast Maarheeze

11

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet


Step 3. Download shipped data from JDE

• Raw data download from JDE

• Transfer into pivot ready

• Shipped quantity per product family

12

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet



Step 4. Determine selection period CV

• Product family

– Select row from pivot

– Determine selection period

– Evaluate effect of elephant orders

13

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet


Row Labels 1008 1009 1010

Altilon 4,266 10,038 3,922

Flash PWF5 250,000 310,760 400,020

Flash PWF6 2,800 1,010,200 251,450

Rebel PC Amber 310 135,090 44,275

Rebel WW Lumiramic 11,500 58,000 30,000



Grand Total 268,876 1,524,088 729,667

Product Family Row Avg Stdev CV Kolom Week Kolom Week

Rebel WW Lumiramic 33 $I$33:$N$33 111,098 151,008 1.36 9 1015 14 1020

Flash PWF5 30 $B$30:$N$30495,145 169,351 0.34 2 1008 14 1020

Flash PWF6 31 $B$31:$N$31678,072 479,877 0.71 2 1008 14 1020

Rebel Hikari 2700K 80CRI T&R 34 $B$34:$N$34 2,000 #DIV/0! 1.00 2 1008 14 1020


Altilon C2 29 $B$29:$N$29 7,918 6,218 0.79 2 1008 14 1020

Rebel PC Amber 32 $B$32:$N$32 94,471 79,168 0.84 2 1008 14 1020

Start week End Week


Step 5. Determine moving average period

• By changing the moving

average in a higher value

demand development of the

coming three months are

taken into account.

• Default value is 2 (outlook of

2 months) - (production lead

time and safety stock levels

are at the right level to

anticipate within a period of 2

months).

14

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet



Step 6. Update average LT & LT variance and

determine service level

• If applicable change the production lead time of the back-end including

transport of Maarheeze including the variance.

• Default value of the service level is 95% -

15

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet

15


Step 7. Copy new ROP values in weekly

ROP sheet

• Copy ROP values into the

weekly ROP calculation

sheet.

16

1. Review product

life-cycle2.


3. Downloa

d shipped

data from JDE

4. Determi

ne selection period

CV

5. Determi

ne moving

average period

6. Update


determine

service level

7. Copy new ROP

values in

weekly ROP sheet



TO

R m

on

thly

RO

P m

ee

tin

g

18

Terms Of Reference SCM monthly review meeting – pilot PWF5

Meeting Frequency Time

Duration Location

Every 3nd Tuesday of the (Ph) month 4 PM (local time Penang) 60 minutes Telephone conference (participant code *****)

Participants SC improvement manager Penang (chair) Planning manager Penang Sr. planning officer (rebel Lumiramic)

Penang Sr. engineering planning officer Penang Integral planner Maarheeze

Objectives Review target settings KPIs incl yield & lead-

times Review target setting ROP Review the balance between the platelet – end

the pump bin distribution Review safety stock targets Review stock levels supermarkets front & back

end Product lifecycle review (product classification)

Input Stock levels front/back -end supermarket Yield Mhz, Png, Sgp Lead-times Mhz, Png, Sgp S&OP sheet KPIs performance overviews Proposal new ROP levels (Sr. engineering planning officer Penang) Future capacity constraints (Lumiramics, or GGI) Output New ROP levels Updated product life cycle table (NPI, Ramp Up, Mature, EOL)

mapping to scheduling method (ROP, on order)

Leading questions KPIs on target? Stock levels supermarket front -end on target level (2 wks of prod

fc(?))? Supply & demand in balance; safety stock -level performance? Platelet distribution in balance with the pump bin availability

/distribution? Sales forecast reliability above 80%? Is everybody agreeing on the new ROP levels?

Agenda KPI development & trend last

period Product lifecycle classification Review platelet distribution Stock development (bin level)

back /front end supermarket incl safety levels

Proposal new ROP levels Any other relevant business

KPIs Confirmed volume performance Maarheeze CLIP Maarheeze, work order (product & bin)

level CRSD, customer requested shipping date

Lumiramics based products Inventory days platelets (current inventory

platelets Penang : daily run rate lumiramic based fc)

Inactive platelets as % of total platelet stock

Basic values Meeting will start and finish on time Stick to the agenda One meeting The appointed employee is responsible for

his/her actions Open and fair discussion Trust each other No reaction = agreeing Be eager to achieve the required result.


RACI – monthly process

19



SC im

prove

men

t man

ager P

enang

Mate

rial m

anage

r Penan

g

Inte

gral p

lanner M

aarheeze

Plate

let p

lanner

Penan

g

Operatio

ns man

ager l

umira

mics

Mhz

Senio

r Busin

ess A

nalyst

Penan

g

WW

supply

pla

nning

man

ager

WW

supply

pla

nner

Purchas

ing m

anag

er Pen

ang

Purchas

ing o

ffice

r Penan

g

KHOR, LEA

N KIM

Monthly process1 Orientation sheet

Update S&OP C I I A R

Update bin distribution A I R I C

2.1 Copy JDE Sales order detail report into Monthly ROP A R

2.2 Verify Shipped table on completeness A R

2.3 Shipped Pivot

Update product family A R

Determine selection period A R

3 Determine average demand period A R

4 ROP calculation

Determine service level A I R

Determine average lead time back-end Mhz I R I A

Determine average lead time variation back-end Mhz I R I A

5 Update Product Life Cycle overview I I R A

Legend:As discussed during WS3 - Penang W20

Legend:



R

A

C

I



Run book monthly ROP meeting

• Meeting will take place at the 3rd Tuesday of the Philips month

• Preparation of the meeting and a pre alignment of the outcome of the

calculation together with Maarheeze is planned at the day before

20


Appendix J Weekly replenishment process


Weekly process

21

1. Import platelet raw data from

JDE in excel2. Update pivot and plot the

stock levels in the tabel

3. Determine

WIP backend

Maarheeze

4. Update in transit sheet

5. Check masterdata

6. Import build plan &

review distribution

7. Load GGI starts

8. Align calculation results with

Mhz


Step 1. Import platelet raw data from JDE

in excel

• Copy download of JDE into the weekly excel

• Don’t change anything at the format

• The next step will be arrange via formulas – to become pivot ready

22

Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.

Determine WIP backen

d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz

Action

1

Action

2



Step 2. Update pivot and plot the stock

levels in the table

• Update pivot

• The values are copied automatically (by macro)

23

Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz

• Be aware that if there is

no bin identification the

stock pieces are not taken

into account.


Step 3. Determine WIP backend Maarheeze

• Maarheeze back-end figures - provided by Adrian at (target) bin level

24

Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz



Step 4. Update in transit sheet

• Based on the information we

have in the in transit sheet

(copy at our share point with

a weekly update).

• Input managed by Adrian

25

Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz


Step 5. Check master data

• All MD as used in the

calculation is collected at

one tab (5 MD).

• Based on those values

calculations will run.

26

Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz


Step 6. Import S&OP & review distribution

27

Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz

• Copy monthly S&OP and translate into weekly demand, bin distribution

included.

S&OP

Product Family 12nc - Maarheeze Lumi - Penang Wknr 1027 1028 1029 1030 1031

Flash PWF5 3222 023 62010 LL60.0163 B330 264,613 264,613 264,613 264,613 264,613

Flash PWF5 3222 023 63010 LL60.0163 B440 617,431 617,431 617,431 617,431 617,431

Rebel WW Lumiramic 3222 023 61200 LUMI.0187 B231 9,540 9,540 9,540 9,540 9,540







Rebel WW Lumiramic 3222 023 61200 LUMI.0187 B262 159 159 159 159 159

Flash PWF6 3222 023 63310 LL60.0134 B222 255,592 255,592 255,592 255,592 255,592

Flash PWF6 3222 023 63410 LL60.0134 B333 766,777 766,777 766,777 766,777 766,777

Flash PWF6 3222 023 63610 LL60.0134 B444 1,150,166 1,150,166 1,150,166 1,150,166 1,150,166

Flash PWF6 3222 023 63510 LL60.0134 B555 383,389 383,389 383,389 383,389 383,389

Jul-10



Step 7. – GGI starts

• Copy GGI starts, planned GGI starts for next week are the expected

consumption 2 weeks later.

28

GGI Starts (history)

Product Family 12nc - Maarheeze Lumi - PenangWknr 1023 1024 1025 1026

Flash PWF5 3222 023 62010 LL60.0163 B330 209600 209600 209600 209,600

Flash PWF5 3222 023 63010 LL60.0163 B440 503040 503040 503,040 503,040

Rebel WW Lumiramic 3222 023 61200 LUMI.0187 B231 0 0 0 0








Flash PWF6 3222 023 63310 LL60.0134 B222 272480 272,480 544,960 230,560

Flash PWF6 3222 023 63410 LL60.0134 B333 1655840 838,400 838,400 691,680

Flash PWF6 3222 023 63610 LL60.0134 B444 544960 1,236,640 964,160 1,048,000

Flash PWF6 3222 023 63510 LL60.0134 B555 272480 419,200 419,200 356,320

Jun-10


Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz

Step 8. Input overview & Calculate replenishment

• Replenishment = ROP – (Pipeline – Usage) rounded to next multiple of 110000

• (G10); 1,672,168 – (200,000+101,519+997,250-115,161) = 488560, rounded: 550000

• If GGI starts are entered in G6, e.g. 10, then G7 will use 10*21840 = 218400, instead of 115,161

• Econ Stock = current pipeline + planned / actual supply – planned / actual usage (actual overwrites

planned)

• (G15); 1,298,769 + 200,000 - 115,161 = 1,383,608 (we now use G11 as actual, not G10 as planned)

• Purpose : to calculate end of period econ stock and project out into the future with MRP build plan and

determine how many weeks we can continue production with the current pipeline stock.

29

If actual is empty

(G8), then G7 is

used

If actual is empty

(G11), then G10

is used. In this

case G11 is

used.



Step 9a. Align calculation results with Mhz

• Step 9a shows the original replenishment quantities

30

1. Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz

Overview Replenishment calculations

Wknr 1027 1028 1029 1030 1031

Flash PWF5 3222 023 62010 LL60.0163 Econ Stock 410,191 575,578 420,965 376,351 441,738

B330 ROP 639,058 639,058 639,058 639,058 639,058

Replenishment 228,867 63,480 218,093 262,707 197,320

Flash PWF5 3222 023 63010 LL60.0163 Econ Stock 852,837 918,975 961,544 894,114 936,683

B440 ROP 1,491,134 1,491,134 1,491,134 1,491,134 1,491,134

Replenishment 638,297 572,159 529,590 597,020 554,451

Rebel WW Lumiramic 3222 023 61200 LUMI.0187 Econ Stock 277,679 258,599 249,058 239,518 229,978

B231 ROP 30,549 30,549 30,549 30,549 30,549

Replenishment 0 0 0 0 0


B232 ROP 59,352 59,352 59,352 59,352 59,352



B241 ROP 91,584 91,584 91,584 91,584 91,584


Jul -10


Rounding Replenishment to Production Batches Jul -10

Product Family 12nc - Maarheeze Lumi - Penang Wknr 1027 1028 1029 1030 1031

Flash PWF5 3222 023 62010 LL60.0163 B330 330,000 110,000 220,000 330,000 220,000

Flash PWF5 3222 023 63010 LL60.0163 B440 660,000 660,000 550,000 660,000 660,000









Flash PWF6 3222 023 63310 LL60.0134 B222 0 0 0 236,000 236,000

Flash PWF6 3222 023 63410 LL60.0134 B333 590,000 708,000 826,000 708,000 826,000

Flash PWF6 3222 023 63610 LL60.0134 B444 1,298,000 1,180,000 1,062,000 1,180,000 1,180,000

Flash PWF6 3222 023 63510 LL60.0134 B555 0 236,000 354,000 472,000 354,000

Step 9b. Align calculation results with Mhz

• Step 9b does the rounding to the nearest production batch sizes

31

1. Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz



Step 9c. Align calculation results with Mhz

• In Step 9c the requested (and rounded) replenishment quantities can be

confirmed (and modified) by Maarheeze.

32

1. Import platelet

raw data from

JDE in excel

2. Update pivot

and plot the

stock levels in

the tabel 3.


d Maarhe

eze

4. Update

in transit sheet5.

Check masterd

ata

6. Import build

plan & review

distribution

7. Load GGI

starts

8. Align calculati

on results

with Mhz

Confirmed Orders MHZ < 1 wk < 2 wks < 3 wks not used

12nc Lumi Wknr 1027 1028 1029 1030 1031

Flash PWF5 3222 023 62010 LL60.0163 B330 430,000 110,000 220,000 330,000 220,000

Flash PWF5 3222 023 63010 LL60.0163 B440 861,000 660,000 550,000 660,000 660,000

Rebel WW Lumiramic not existing LUMI.0187 B221 0 0 0 0 0

Rebel WW Lumiramic not existing LUMI.0187 B222 0 0 0 0 0









Flash PWF6 3222 023 63310 LL60.0134 B222 0 0 0 236,000 236,000

Flash PWF6 3222 023 63410 LL60.0134 B333 948,000 708,000 826,000 708,000 826,000

Flash PWF6 3222 023 63610 LL60.0134 B444 1,300,000 1,180,000 1,062,000 1,180,000 1,180,000

Flash PWF6 3222 023 63510 LL60.0134 B555 356,000 236,000 354,000 472,000 354,000

Jul -10


TOR weekly ROP meeting

44

Terms Of Reference SCM weekly review meeting

Meeting Frequency Time Duration Location

Weekly Monday - 5 PM (local time Penang) 60 minutes Telephone conference (participant code *****)

Participants SC improvement manager Penang

(chair) Planning manager Penang Sr. planning officer (rebel Lumiramic)

Penang Sr. engineering planning officer

Penang Integral planner Maarheeze

Objectives Discuss result production Penang, Maarheeze & Singapore o f

the last week by reviewing defined KPIs . Based on KPI review determine improvement areas including

corrective actions Set priorities for quality & quantity related topics Discuss progress running actions linked to this meeting

Input Action list SCM review meeting Weekly platelet planning sheet

previous week KPIs performance overviews Output Updated weekly platelet planning

sheet Updated action list SCM review

meeting

Leading questions Winning strikes of last week Attention areas of last week Taken actions and results Actions taken to prevent our organization for the bleeders of today Short term planned actions

Agenda Results last week & results next

week o Maarheeze o Penang

Stock levels back-end Maarheeze Outcome weekly platelet planning

sheet Review main actions Action list update Any other relevant business

KPIs Confirmed volume performance Maarheeze CLIP Maarheeze, work order (product & bin)

level CRSD, customer requested shipping date

Lumiramics based products Inventory days platelets (current inventory

platelets Penang : daily run rate lumiramic based fc)

Inactive platelets as % of total platelet stock

Basic values Meeting will start and finish on time Stick to the agenda One meeting The appointed employee is responsible fo r his/her

actions Open and fair discussion Trust each other No reaction = agreeing Be eager to achieve the required result.



RACI - weekly

45



SC im

prove

men

t man

ager P

enang

Mate

rial m

anage

r Penan

g

Inte

gral p

lanner M

aarheeze

Plate

let p

lanner

Penan

g

Operatio

ns man

ager l

umira

mics

Mhz

Senio

r Busin

ess A

nalyst

Penan

g

WW

supply

pla

nning

man

ager

WW

supply

pla

nner

Purchas

ing m

anag

er Pen

ang

Purchas

ing o

ffice

r Penan

g

KHOR, LEA

N KIM

Weekly process0 Prepare KPI reports

Penang part A R

Maarheeze part R A

1a Copy JDE reports Stock Aging report into weekly ROP A R

1b Verify Platelet stock table on completeness A R

1 Copy Stock take values into table A R

2 Determine WIP back-end Mhz R A

3 In transit

Update shipped goods R A

Copy JDE report received goods A R

4 Copy ROP from monthly to weekly ROP sheet A R

5 MD update A R

6 Copy GGI schedule Penang into weekly ROP A R

7 Copy # of GGI starts into weekly ROP A R

8 Verify GGI history sheet on completeness A R

9a Verify Calc replenishment on completeness /rightness A R

9b Verify Production Orders MHZ on completeness /rightness A R

Legend:As discussed during WS3 - Penang W20

Legend:



R

A

C

I


Run book weekly ROP meeting

• Meeting will take place at every Thursday (5pm Penang local time)

• Preparation and pre alignment planned at Wednesday (Maarheeze)

/Thursday Morning (Penang)

46


Appendix K Tutorial Supply chain control models

PHILIPS LIGHTING, BU LUMILEDS

Tutorial Re-Order-Point

calculation platelets

Penang Planning method for lumiramic based products

Penang and Maarheeze

[July 2010]

This document will help you to understand the re-Order-Point calculation method including the

independencies between the organizations involved.


Purpose of this document

This document will explain how to use and maintain the Re-Order-Point calculation excel sheets.

Re-Order-Point process

Overall context

The figure below shows the current lumiramics business process at high level. The numbers in the figure

(described in the text below) highlights the most critical process steps. Secondly, the black line indicates

the situation before the project the red line indicates the situation at the end of the project. The design

of the ROP calculation will cover the back-end processes of Maarheeze including the GGI starts. The GGI

starts is the main trigger for starting production batches in the back-end of Maarheeze. The front-end of

Maarheeze is managed based on the Kanban principle. The level of Kanban tickets in the front-end of

Maarheeze is directly linked with the ROP calculation.

ROP GGI/Saber/Assy&Test

1. In this phase of the to-be we’ll keep the GGI starts planning the same as in the as-is (MRP build-

plan GGI),

2. The same for the EPI starts (so no changes here).

Build plan GGI

Determine production batches on pump-bin level

(checking pump-bin and platelet bin availabilities, finding combinations wafer types, pump-bins,


platelets, among competing requirements other products)

will be investigated in a broader scope (not just lumiramics) and supported by a LP tool.

Replenish platelet bank

3. ROP on platelet bank in Penang

ROP calculation based on forecast and bin distribution

ROP execution based on GGI production orders accumulated per internal partnr / bin, platelet

stock, WIP in Maarheeze and in-transit

4. Replenish the pre-grinded wafer bank

(using Kanban, as product costs in front-end are relatively low)

5. Anticipate on started GGI production orders from (1)

data flow from GGI starts to platelet bank (as future consumption)

Focus areas

As mentioned in the introduction of this chapter we will limited ourselves to the supply of lumiramics

from Maarheeze to Penang. In the IDEF0 (Integration Definition for Function Modeling) diagram below

the overall stock calculation process is decrypted in a structured matter.

Capacity

constraints

Product-life-

cycle

TITLE:NODE: NO.: 1To be

A0SADT Lumileds /lumiramics

A3

WIP back-end

Maarheeze

A1

Monthly ROP

calculation

A4

Generate In-

transit overview

A2

Weekly

replenishment

S&OP

Shipments

Service levelMAT

Lead-time platelets

WIP Maarheeze

In-transit

Consolidated stock data overview per bin

Yield back-end Maarheeze

Stock pre grinded wafers

Carrier lead-time

Completions to inventory

Consolidated stock data

overview

GGI start

Period


GI /Invoice

MES

Supermarket front-end

Maarheeze

Stock data overview

MES

A5

Calculate platelet

stock position

Png

ROP

values

WIP Back-end

Mhz

Confirmed

plan

Requested

plan

In-transit

overview

Prod family table


A1. The process design starts with a Re-Order-Point calculation, to be executed on a monthly frequency.

As input for the Monthly ROP calculation the following elements are necessary;

The platelet distribution,

The platelet orientation out of the S&OP,

Shipments of the last 6 months for determination of the variance,

The transport lead-time and his variation and

A product family table Control elements are;

Selected period of customer shipments (for variance determination)

The MAT (moving average total) which incorporate the demand development for the coming periods

The product lifecycle, the ROP calculation method is designed for all phases in the product lifecycle but has a best fit in the maturity phase of a product.

Service level, with the service level indicator you will infect directly the level of safety stocks. The higher the services level the higher the safety stock.

Output element;

The ROP value per product family.

A2. Replenishment process

Input elements

Work-In-Process (WIP) back-end Maarheeze, all back-end started production batches but not

yet completed.

In-transit, production batches which has left Maarheeze but without a goods receipt in Penang.

Consolidated stock overview per bin, all lumiramics booked on stock in Penang

Yield back-end Maarheeze, on product level

GGI starts, number of blue bin batches planned for startup in the GGI process. Based on this

information we know almost exactly the consumption in two weeks from now. See also picture

below;


Control elements

ROP values as set in the monthly ROP calculation (as described at A1)

Capacity constraints lumiramic based products – max capacity limits at Maarheeze.

Output element

Requested plan

Confirmed plan

A3. WIP back-end Maarheeze

Input elements

GI/Invoice – for measuring output of the back-end of Maarheeze the SAP Invoice or Goods Issue

information is leading. Without any SAP Invoice no delivery towards Penang.

Supermarket front-end Maarheeze, the basic material for the final product of Maarheeze.

Without stock in the front-end supermarket (on product bin level) no start up of production

batches in the back-end is possible.

Control elements

Requested plan, the first parameter is set by the initial request out of the replenishment

calculation. This calculation is mainly based on the boundaries as set in the ROP calculation and

the GGI starts.

Confirmed plan, based on the request Maarheeze will plan the back-end with as input the

supermarket (available raw materials and the available capacity) a production plan is

communicated towards Penang. This (confirmed) plan is the benchmark of the back-end

production of Maarheeze.

2010-05-02 2010-06-06

5-9 5-16 5-23 5-30

Singapore5-3 - 5-5

GGI /Saber

planning

2010-05-02 2010-06-06

5-9 5-16 5-23 5-30

Penang

2010-05-02 2010-06-06

5-9 5-16 5-23 5-30

Maarheeze5-6 - 5-7

B-E

planning

5-10 - 5-21

Back-end production

5-21 - 5-25

Transport

5-10 - 5-23

GGI /Saber production (throughput time; 3 + 7 days)

5-23 - 5-25

Transport

GGI /Saber is leading

Bin distribution based on fc

5-30 - 6-6

Start production

EXAMPLE


A4. Generate In-transit overview

Input elements

WIP (work-in-process) back-end Maarheeze, the batches of the final products of the back-end of

Maarheeze are transferred to the in-transit sheet.

Stock data overview, based on the match at batch level between the goods receipt entries

(physical available stock of Penang) and the output of the back-end Maarheeze the final

overview can be generated.

Control elements

Carrier lead-time is the control element in this process, the longer the lead-time the bigger the

stock quantities in-transit.

A5. Calculate platelet stock position Penang

Input elements

In-transit overview, output of the process as descript at A4.

Completions to inventory, all production batches which are booked in.

Control elements

ROP values, as described in A1 the ROP boundaries are set in the calculation. Physical and

average available stock level in Penang according the model is the safety stock and approx 50%

of the demand during lead-time.


Monthly Re-Order-Point calculation

Model setup

Step 1 – review product life cycle

On final product level, meaning one coefficient of variance per final product is valid for multiple

(lumiramic) bins. In the table below the threshold values are mentioned, based on the threshold value

the calculation can be followed freely without too much manual intervention.

1. Review product life-

cycle

2. Prepare orientation

sheet

3. Download shipped data

from JDE

4. Determine selection period CV

5. Determine moving average period

6. Update average LT &

LT var and determine

service level

7. Copy new ROP values in weekly ROP sheet

Product life cycle Characteristics Threshold value Planning method

New product

introduction

Slow volumes to start

and demand has to be

created

CV > 1 Re-order-point &

manual intervention

Ramp up (growth) Sales volume

increases significantly


manual intervention

Maturity Sales volume peaks

and market saturation

is reached

CV < 1 Re-order-point

calculation

Decline (End of

Live)

Sales volume decline

or stabilize


manual intervention


Step 2 – prepare orientation sheet

Based on the S&OP a lumiramics supply plan is generated. The yellow fields are changeable (period,

total need and distribution) the rest of the fields are filled with formulas. None yellow fields are write

protected.

Step 3 – download shipped data from JDE

Download the standard report from JD Edwards (R554211C) – no changes in format, deleting columns or

whatever.

At the next modifications are made, empty columns are deleted. Column Q and R are calculated fields

(Wknr based on actual shipped date [column M] & Prod fam based on 2nd item number [column H]), be

aware in case of many shipments (more than 27 k) the formulas must be copied down (A ..R).

Regarding the product family as mentioned in column R – the 2nd item number of column H is linked

with a table at tab.5

Orientation based on monthly S&OP update

Phosphor Product Line Product Family Lumi nmbr Penang 12nc Maarheeze Bin Distribution Jul-10 Aug-10 Sep-10 Oct-10

LUMIRAMIC Luxeon Flash Flash PWF5 4,410,220 2,566,173 2,688,853 3,931,685

LL60.0163 3222 023 62010 330 30% 1,323,066 769,852 806,656 1,179,505

LL60.0163 3222 023 63010 440 70% 3,087,154 1,796,321 1,882,197 2,752,179

Luxeon Flash Flash PWF6 12,779,618 8,456,514 8,036,564 10,555,193

LL60.0134 3222 023 63310 222 5% 638,981 422,826 401,828 527,760

LL60.0134 3222 023 63410 333 35% 4,472,866 2,959,780 2,812,798 3,694,317

LL60.0134 3222 023 63610 444 45% 5,750,828 3,805,431 3,616,454 4,749,837

LL60.0134 3222 023 63510 555 15% 1,916,943 1,268,477 1,205,485 1,583,279

Luxeon Rebel Rebel WW Lumiramic 495,854 388,353 459,362 573,198

LL60.0187 3222 023 61200 221 0% - - - -

LL60.0187 3222 023 61200 222 0% - - - -

LL60.0187 3222 023 61200 231 10% 47,706 37,363 44,195 55,147

LL60.0187 3222 023 61200 232 19% 92,684 72,590 85,863 107,142

LL60.0187 3222 023 61200 241 29% 143,019 112,012 132,493 165,327

LL60.0187 3222 023 61200 242 26% 130,770 102,419 121,146 151,168

LL60.0187 3222 023 61200 251 11% 54,301 42,529 50,305 62,772

LL60.0187 3222 023 61200 252 3% 16,563 12,972 15,344 19,147

LL60.0187 3222 023 61200 261 2% 10,017 7,846 9,280 11,580

LL60.0187 3222 023 61200 262 0% 793 621 735 917

R554211C Philips LumiLeds Lighting

Sales Order Detail Report

Order Number Or Ty Order Co Line Number Hd CD Request Date Sched Pick

363961 SI 10 50 2009-12-15 ########

363961 SI 10 60 2010-01-08 ########

363961 SI 10 70 2010-02-12 ########

363961 SI 10 80 2010-03-12 ########


Step 4 – determine selection period CV

The pivot will refresh automatically after entering the tab. Check after refresh if the total data source is

selected via pivotTable Tools| change data source. In case an unknown family name is popping up as

part of the row labels; the product family table at tab 5 is outdated. Actions to take- update product

family table at tab 5 and refresh the pivot.

2nd step is to determine the CV value (coefficient of variance) is the standard deviation divided by the

average. In column Row [B] you have to select the right row of the product family in the pivot. For

selecting the requested time window use the yellow columns at the right side [G&I].

R554211C

Order

Number Or Ty Order Co

Line

Number

Request

Date

Sched

Pick

Order

Date

363961 SI 10 50 40162 40221 40162

363961 SI 10 60 40186 40221 40162

363961 SI 10 70 40221 40221 40162

363961 SI 10 80 40249 40249 40162


Row Labels 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 Grand Total

Altilon 4,266 10,038 3,922 12,436 30 18,396 4,430 35 7,943 9,544 14,304 1,102 16,493 6,283 109,222

Flash PWF5 250,000 310,760 400,020 885,150 585,000 580,000 455,000 410,000 490,900 375,000 538,050 447,000 710,000 265,000 6,701,880

Flash PWF6 2,800 1,010,200 251,450 650,000 320,100 960,000 1,102,000 333,150 500,150 1,045,000 1,600,020 50,064 990,000 1,312,000 10,126,934

Rebel PC Amber 310 135,090 44,275 59,070 75,000 112,150 109,020 116,304 276,040 100,010 194,100 6,700 55 106,095 1,334,219

Rebel WW Lumiramic 11,500 58,000 30,000 35,000 32,500 9,050 39,040 410,000 32,500 124,050 22,000 39,000 98,000 940,640

Platelet, Hikari 2700K 80CRI 2,000 450 2,450

Platelet, Hikari 3000K 80CRI 2,000 2,000

Grand Total 268,876 1,524,088 729,667 1,641,656 1,012,630 1,670,546 1,679,500 898,529 1,685,033 1,562,054 2,470,524 526,866 1,759,548 1,787,828 19,217,345

Product Family Row Avg Stdev CV Kolom Week Kolom Week

Rebel WW Lumiramic 33 $I$33:$N$33 111,098 151,008 1.36 9 1015 14 1020

Flash PWF5 30 $B$30:$N$30 495,145 169,351 0.34 2 1008 14 1020

Flash PWF6 31 $B$31:$N$31 678,072 479,877 0.71 2 1008 14 1020

Rebel Hikari 2500K 80CRI T&R - NA NA NA 1.00 2 1008 14 1020










RebelHikari 4000K 70CRI T&R - NA NA NA 1.00 2 1008 14 1020




Altilon C2 29 $B$29:$N$29 7,918 6,218 0.79 2 1008 14 1020

Rebel PC Amber 32 $B$32:$N$32 94,471 79,168 0.84 2 1008 14 1020

Start week End Week


Step 5 – determine moving average period

Use column MAT [Q] for changing the moving average total, the MAT takes care about the future

changes in the S&OP. The default value for the MAT is 2, meaning the forecast of the 2 coming months

are taken into account. In case the value is set at 3 the coming 3 months are taken into account,

etcetera. The MAT is used as weekly demand in the ROP calculation.

Step 6 – parameters to set for calculating the economical stock levels (ROP) are in (the yellow marked

fields) column S, T, U, V, W.

Column S: supply variance back-end Maarheeze measured on planning reliability.

Column T: supply variance back-end maarheeze measured on quantity reliability.

Column U: lead-time back-end Maarheeze in weeks.

Column V: lead-time GGI/Saber Singapore in weeks.

Column W: requested service level as agreed at management level between Maarheeze and

Penang.

Column S, T, U requires input of the integral planner of Maarheeze. Column V is the responsibility of the

platelet planner of Penang. Column W is the responsibility of the demand manager Penang.

The ROP is the sum of the safety stock and the demand during lead-time.

Column A – G are part the standard format repeating at at tab.

Column H, ROP calculation = I + J

Column I, Demand during leadtime = K (weekly demand) * L (LT1 = U + transport lead-time of 1 week).

Column J, Safety stock = P * SQRT(M*(O*K)^2+K^2*(N*M)^2)

M refer to the demand lead-time of GGI/Saber versus LT1 with a minimum of 1 week

(replenishment lead-time)

N refer to the supply variance = SQRT(S^2+T^2)

O refer to the demand variation as calculated at tab 2.3

P refer to the service level as indicated in column W

Phosphor Product Line Product Family Lumi nmbr Penang12nc Maarheeze Bin Distribution MAT Period 1 MAT period 1 Period 2 MAT period 2 Period 3 MAT period 3LUMIRAMIC Luxeon Flash Flash PWF5 2 Jul-10 / Aug-10 761,794 Aug-10 / Sep-10 656,878 Sep-10 / Oct-10 729,275

LL60.0163 3222 023 62010 330 30% 228,538 197,063 218,783

LL60.0163 3222 023 63010 440 70% 533,256 459,815 510,493

Luxeon Flash Flash PWF6 2 Jul-10 / Aug-10 2,335,026 Aug-10 / Sep-10 2,061,635 Sep-10 / Oct-10 2,060,090

LL60.0134 3222 023 63310 222 5% 116,751 103,082 103,004

LL60.0134 3222 023 63410 333 35% 817,259 721,572 721,031

LL60.0134 3222 023 63610 444 45% 1,050,762 927,736 927,040

LL60.0134 3222 023 63510 555 15% 350,254 309,245 309,013

Moving average total

ROP calculation sheet E F G H I J K L M N O P Q R S T U V W

(MAT)

Phosphor Product Line Product Family Lumi nmbr Penang 12nc Maarheeze Bin ROP Dem LT. Safety stock Weekly dem. LT 1 LT 2 Suppl var. Dem var. Z Safety[wks] Avg stock [wks] Supply var. t Supply var.q LT BE LT GGI Service l.

LUMIRAMICLuxeon Flash Flash PWF5

LL60.0163 3222 023 62010 330 766,037 576,569 189,468 228,538 2.52 1.00 0.35 0.36 1.64 0.83 1.33 0.33 0.13 1.52 1.71 95%

LL60.0163 3222 023 63010 440 1,845,187 1,374,276 470,911 533,256 2.58 1.00 0.42 0.36 1.64 0.88 1.38 0.40 0.13 1.58 1.71 95%

Luxeon Flash Flash PWF6

LL60.0134 3222 023 63310 222 482,663 331,740 150,922 116,751 2.84 1.13 0.45 0.61 1.64 1.29 1.79 0.40 0.20 1.84 1.71 95%

LL60.0134 3222 023 63410 333 3,716,094 2,317,513 1,398,581 817,259 2.84 1.12 0.76 0.61 1.64 1.71 2.21 0.73 0.20 1.84 1.71 95%

LL60.0134 3222 023 63610 444 4,592,884 3,068,224 1,524,660 1,050,762 2.92 1.21 0.52 0.61 1.64 1.45 1.95 0.48 0.20 1.92 1.71 95%

LL60.0134 3222 023 63510 555 1,548,109 1,049,761 498,348 350,254 3.00 1.28 0.46 0.61 1.64 1.42 1.92 0.41 0.20 2.00 1.71 95%


Q is a calculation field = J/K representing the safety stock levels in terms of weeks of demand as

presented in column K.

R present the number of weeks physical available stock in Penang = the safety stock + ½ week.

The ½ week representing the average stock level due to consumption in a replenishment period

of 1 week.

Step 7 – copy new ROP levels in weekly ROP (replenishment) sheet

Copy column *H+ ‘ROP’ to the weekly replenishment sheet and save close the monthly_lumiramics.xls.

Maintain procedure

The Monthly_lumiramics.xls is a protected sheet; the password of each tab is drop. In the protected

mode users are only allowed to modify the yellow colored fields.

The standard fields like article number, bin distribution, month and etcetera are all linked to the

corresponding fields at tab 1 [Orientation sheet].

In case of adding or changing information at material master level, start at tab1. In case of a new

commercial type don’t forget to update the product family table at tab5.

In case of adding a new product to the ROP calculation sheet, start at tab1 with insert new row. Repeat

inserting new row (same position) at tab2.2, tab3, tab4 and copy the formulas from the row above.

FAQ

1. Q: Replenishment is arriving too late at Penang. Q: Increase transport lead-time and/or

transport lead-time variation at tab4 [column K and M].

2. Q: Request of weekly replenishment is more than we can deliver. A: Compare demand at

tab1.[orientation] & tab3.[avg.dem] with the requested quantities based on the GGI starts in the

weekly replenishment sheet. If the deviation is more than 10% re-consider the ROP by a manual

update of the CV value at tab 2.3 (the higher the CV value the higher the ROP value). And copy

the revised ROP value to the replenishment sheet. Communicate the issue to the S&OP

/Demand planner.

3. …


Weekly Replenishment calculation

During the weekly replenishment cycle we determine the actual quantities of platelets to be replenished

by Maarheeze. The timing is chosen close to the end of the week (N-1) in order to get the best estimate

of the begin-of-the-week N status with respect to WIP, In-Transit and On-hand Inventory. This gives us

the economical stock at the start of week N. The planned GGI starts for week N give us the demand for

week N. The required replenishment for week N is easily determined by ROP – economical stock +

demand week N (1000 – 800 + 300 = 500; 800 – 1000 + 300 = 100)

The main efforts in the weekly cycle are to get all the inputs and in a synchronized way (in order to avoid

double counting). Typical inputs: Stock take of platelets in Penang, Work in process in Maarheeze, In-

Transit between Maarheeze and Penang, S&OP monthly quantities and the GGI starts for next week.

With this input the replenishments will be calculated automatically. The requested quantities are given

to Maarheeze, and Maarheeze will respond with the confirmed quantities based on their current

capabilities (capacity, front-end supermarket, yield situation, etc.)

Model setup

1. Import platelet raw

data from JDE in excel

2. Update pivot and

plot the stock levels in the

tabel

3. Determine WIP backend Maarheeze

4. Update in transit sheet

5. Check masterdata

6. Import build plan &

review distribution

7. Load GGI starts

8. Align calculation results with

Mhz


These process steps are supported by an excel workbook with the following sheets:

Name Description

Action list

Instructions

1a Platelets raw data Input from JDE

1b Platelet stock Transform 1a

1c Stock Pivot Pivot of 1b

1 Stock take Select 1c into one column

2 WIP Input from MHZ

3 In Transit InTransit

4 ROPs Input Monthly ROPS

5 MD Master Data

6 SnOP Input Constrained Lumiramics S&OP

7 GGI Starts Input Planned GGI starts

7a Backlog Input for GGI backlog

8 Replmts Replenishment Calculation and input summary

9a Calc Replenishment Replenishment details

9b Production Orders MHZ Replenishment rounded

9c Confirmed Orders MHZ Confirmed replenishment

8 StockDays Auxiliary sheet for highlighting 9b & 9c

Summary On-hand projection (still in development)

WIP Reports Input from Sorting (LL60 to Lumi)

Step 1 Import platelet bin raw data from JDE into Excel

The standard JDE report “Stock Aging Report By BIN (excel Format) with Sales Cat Code” is loaded into

sheet “1a Platelets raw data”. This should be the exact standard report without any modifications. The

reason for this: we will transform this report into the required bin information automatically. However,

in order to do this we need to rely on a stable (not likely to change) report layout.

The transformation consists of the following steps:

Eliminate the empty columns, otherwise we cannot make a pivottable. This is done in sheet “1b

Platelet stock”.

In this sheet we also use the tLumi function in column B. This transforms the LL60 names to

LUMI, in order to simplify the summation of the stock take in the pivot table (except for Flash

LL60.0134 and LL60.0163)

In sheet “1c Stock Pivot” the data of “1b Platelet stock” is summarized in a pivot table by Item

number and by bin.


Finally, in sheet “1 Stock take” the content of the pivot table is transferred to the planning week

column. When opening this sheet the pivot table “1c Stock Pivot” is automatically refreshed and

the content is copied to the planning week automatically. The macro which enables this

functionality is located on the sheet tab (right-click the sheet tab)

Step 2 Determine WIP backend Maarheeze

Based on the input from Maarheeze we fill-in the WIP on the “2 WIP” sheet. This is the work in process

of the backend in Maarheeze.


Step 3 Update in-transit sheet

When platelet shipments leave Maarheeze the shipment information is logged into the “Lumiramics-

Intransit” workbook. Before updating the intransit information on the “3 Intransit” the “Lumiramics-

Intranist” workbook needs to be updated with the JDE report “Completions to Inventory - Transaction

Details”. The JDE report indicates the shipments which have arrived and have been added to the

warehouse. By comparing the JDE report with the Lumiramics-Intransit workbook, the actual shipments

still in transit become visible. This process is also supported by excel-macros, in order to minimize the

manual efforts.

After updating the Lumiramics-Intransit workbook the information can be copied into the “3 Intransit”

sheet.


Step 4 Copy the ROPs from the monthly ROP calculation

Every month we review and recalculate the Reorder Points, as the business is continuously changing.

The newly calculated ROPs need to be transferred to the weekly workbook, to these may be used in the

weekly execution to determine the replenishment quantities

Step 5 Check the master data

The master data is limited to the GGI batch sizes, the GGI/Saber yield, the lumiramics batch sizes, the

platelet distribution and the product family. Typically this data will not change much, but sometimes

adjustments are required.


Step 6 Import SnOP and review distribution

The replenishment requirements of the platelets will closely follow the GGI starts. Also, in order to

anticipate on potential volume and/or mix changes the monthly S&OP at the product family is added. In

order to support the weekly process at bin level, we need to transform the monthly product family

quantities in to weekly bin-level quantities. We do this by dividing by the number of weeks in the month

according to the Philips calendar and by using the bin-distribution.

Typically the bin-distribution should be the same as used for the monthly ROP calculation. It can be

derived from the EPI bin distribution but it should also include the availability of pump-bins for the

product families.


Step 7 Load GGI starts

The planned GGI starts for week N + 1 will be the best forecast for the platelet consumption in week

N+3, under the assumption that it takes GGI/Saber (in Singapore) 2 weeks to arrive at Platelet Attach in

Penang. This time window gives the backend of Maarheeze to respond to the replenishment quantities


Step 7a Load the backlog

In case of backlog, we need to capture this separately, otherwise it is assumed to be netted in the stock

take. The backlog will also be calculated in sheet “8 Replmts”, but the actual backlog will override the

result of the calculation.

Step 8 Replenishment calculation

In sheet “8 replmts” the replenishment calculation is done. Also, all the inputs are summarized per

product family / bin in order to make the calculation result transparent.


Let’s assume we are at the end of week 1028 and planning for week 1029. The pipeline stock can be

found in T13:T15 (taken from “1 Stock take”, “2 WIP”, “3 InTransit”). The monthly calculated ROP can be

found in T10 (from “4 ROPs”). The GGI build plan can be found in T7 (from “7 GGI Starts”) and the

Backlog in T6 (from “7a Backlog”). If the Backlog is specified in “7a Backlog” then this value is taken,

otherwise the value is calculated. The simplified version of the formula is: MAX(0; R7-T15-T14), so if the

GGI starts in week 1027 (R7) are larger than the current stock and intransit, then the remainder is

considered as backlog.

In T8 we take the GGI starts (from T7) if specified, otherwise we take the S&OP value (from “6 SnOP”).

Now we have all the inputs for the replenishment calculation: the current pipeline (T13:T15), the

expected demand (T8), the backlog (T6) and the ROP (T10). The formula is: T10-[SUM(T13:T15))-T8-T6],

so first we determine the remaining economical stock: current pipeline minus expected demand minus

backlog. This result we compare with the ROP (T10) and we reorder the difference (to order up to the

reorder level. The actual formula also does the rounding to the nearest production batch. So, the result

in T11 is the rounded replenishment request.

In T12 the confirmed quantity will be shown, this is facilitated by sheet “9c Confirmed Orders MHZ” and

explained below.

In order to understand how well the ROP replenishment is working we need to calculate to KPIs:

calculate the number of weeks economical stock in the pipeline and the quantity stock on-hand. The

first will indicate whether the ROP is still at the correct level, the second KPI will indicate whether we

have enough inventory to avoid line stoppages.

In T16 we calculate the economical stock at the end of the week: = SUM(T13:T15) + T12 - T8 - T6. So,

basically the economical stock at the beginning of the week + the confirmed supply – GGI starts –

backlog. In case the confirmed supply is not available, the calculated replenishment quantity is given

(T11). Also, if the actual information is not available (so SUM(T13;T15) = 0) then the end stock of the

previous week will be used (S16).

The quantities of row 16 are translated into week of future demand. We are using the numbers of row 8

(GGI Starts, S&OP) as input. The average demand of the first 4 weeks will be used to calculate the

number of weeks stock. The results can be found in row 17 (T17).

The on-hand projection of the stock can be found in row 18. The formula in T18 is SUM(T14:T15) - R8 -

T6, so the current on-hand plus the in-transit minus the GGI Starts of 2 weeks earlier, minus the backlog.

This on-hand projection (T18) should be greater than 0, otherwise (part of) the platelet/attach

production steps may come to standstill due to platelet shortages.

For all the product families / bins the above information has been compiled in sheet “8 Replmnts”. The

drop-down at the upper left hand corner allows to quickly access the required product family / bin.


The details of the replenishment are given in sheets “9a Calc Replenishment”, “9b Production Orders

MHZ”, and “9c Confirmed Orders MHZ”. This will be explained in the next section.

Step 9 review the replenishment proposals and align with Maarheeze

The replenishment calculations are given in sheets “9a Calc Replenishment”. This sheet is for

information purposes only.

In sheet “9b Production Orders MHZ” the replenishments are rounded to the production batches in

Maarheeze. Also, highlighting is used to indicate priorities. The highlighting is related to the economical

stock in weeks (row 17 in “8 Replmts”). This sheet is for information purposes only.

Sheet “9c Confirmed Orders MHZ” is essentially a copy of the previous sheet, and it allows the input of

the confirmed quantities. The calculated values should be replaced with the confirmed quanties. The


highlighting will change color when the confirmed quanties are different and the calculated weeks of

economical stock fall into a different category (< 1 wk is red, < 2 wks is orange, < 3wks is yellow, >= wks

is no highlighting, not used is blue)

Step 10 Summary of on-hand inventory projection

The “Summary” sheet shows the on-hand inventory projection. This sheet has not been finalized yet.

Maintain procedure

In this section we will describe the steps necessary to modify the sheet, e.g. when new product families

/bins have to be added to the sheet. In the following sections we will give an overview of the key

assumptions which are important to understand why certain things cannot or should not be done. Also,

we describe the macros so it becomes clear what these will do.


Adding

The key assumptions for modifying the Weekly_Lumiramics.xlsm workbook are listed below.

• All of the sheets refer to “1 Stock take” for the first 4 columns

• New products become automatically visible after running the AddProduct macro

• This creates consistency but also limitations: moving products is prohibited

• This setup is chosen to minimize the maintenance overhead, but limits the flexibility somewhat

• 1 Stock take (continued)

– The first 4 columns (product family, 12nc MHZ, LUMI Penang, bin) are copied to all the

subsequent sheets. So this is the "master" for all item numbers.

– Adding new item numbers should be done here, and only here.

– The macro AddProduct can be used to create the referring links on all other sheets and

copy the relevant blocks of information

– OnActivate: automatically update the stock take pivot and transfer the values to the

stock take list in week N+1 The macro contains some hard-coded exceptions:

• rows 10-17, for Rebel WW also add the LUMI.0079 to the stock per bin

• rows 22-35, for Hikari only count the total stock (not per bin)

– In case the item numbers move to different rows (see general remark: this should be

avoided), the macro has no way of knowing this, as it goes by row number, and

therefore it needs to be changed to refer to the correct rows

• 6 SnOP

– The input is taken from the monthly ROP (Orientation / SnOP)

– The S&OP is at the product family level, when new products are added the division

among the bins should be added by hand.

– The first row of the first product/bin should not be moved

(because the macro assumes the first row is at row 41)

• 8 Replmts

– This sheet summarizes per product / bin all the input and calculates the replenishment


– As it has many references to all the input sheets, we have a separate macro

"Add8Replmts" to update this sheet

– It copies the range of 12 rows and 6 columns and updates the references, then

copy/paste column 6 towards the end of December

• 9a Calc Replenishment

– Details the Replenishment calculation

– Adding new product is part of the AddProduct macro and in a separate macro:

CalcRepl9a

• 9b Production Orders MHZ

– Rounds the Replenishment requests to MHZ batch sizes

– It uses conditional highlighting to indicate urgency.

– The values used are in sheet "8 StockDays" In order for the highlighting to work with

values on a different sheet we need to use named ranges: StDays

– OnActivate: to move the highlighting to week N and N+1


ProdOrders9b

• 9c Confirmed Orders MHZ

– This sheet shows the values from the previous sheet (9b Production Orders MHZ) and

allows this values/formulas to be overtyped by the confirmed order quantity

– It uses conditional highlighting to indicate urgency.

– if the formula is replaced by a confirmed value, this value will show in the "8 Replmts"

sheet

– OnActivate: to move the highlighting to week N and N+1


ConfirmedProdOrders9c

• 8 StockDays

– This sheet support the highlighting on sheets 9b and 9c.

– The values (economical stock in weeks) are copied from the "8 Replmts" sheet



UpdateStockDays8

– Manually copying the rows down will not work, because then the reference to "8

Replmnts" will be off

• Macro’s

– AddProduct

• For each sheet check with “1 Stock take” and add the missing references to the

stock take sheet (if any). Also update the drop-down list (on “8 Replmts”) to

include all the product families / bins.

– Add8Replmts

• Copy the first product / bin range, paste for the missing products, and update all

references

– CalcRepl9a


references.

– ProdOrders9b


references.

– ConfirmedProdOrders9c


references.

– UpdateStockDays8


references.

– UpdateSheet1

• Copy the product / bins from sheet 1 Stock take and redefine the named range

“ProdFam”

Deleting


It is not recommended to delete rows / columns, because this may prevent the macros from working

correctly. We propose to work with a new file every year, and before starting the operational use, we

can clean up the sheets. We need to make sure that the data in the input sheets is cleaned up correctly.


Kanban calculation front end Maarheeze

Model setup

Number of kanban tickets is based on;

MAT (moving average total – copy from monthly ROP) per product

Lead-time frontend as measured in the MES of Maarheeze

Service level (Z) = 99%

Standard deviation = MAT * Coefficient of variance (copy from monthly ROP)

Production batch size Maarheeze

Formula for calculating the number of kanban tickets per product:

((MAT*lead-time) + (Z * Standard deviation))/ Production batch size

Back-end MhzFront-end Mhz

FIFO

OXOX

FIFO

S&OP Lumiramics

Lumiramics

production Penang

Monthly ROP Weekly

replenishment


Screenshot of the calculation model:

Yellow colored fields are adjustable.

The kanban calculation activities are based on the monthly ROP output. Every month the kanban tickets

of the front-end Maarheeze will be updated.

Last updated; June 2010

Product Family Lumi nmbr Png 12nc Maarheeze Bin D CV LT Q Z STD D*LT Safety stock Totaal

Luxeon Flash5 LL60.0163 3222 023 62010 330 228,538 0.34 2.94 110,000 99% 77,703 7.0 2.0 9.0

Luxeon Flash5 LL60.0163 3222 023 63010 440 533,256 0.34 2.53 110,000 99% 181,307 13.0 4.0 17.0

Luxeon Flash6 LL60.0134 3222 023 63310 222 116,751 0.71 2.34 117,000 99% 82,893 3.0 2.0 5.0

Luxeon Flash6 LL60.0134 3222 023 63410 333 817,259 0.71 2.10 117,000 99% 580,254 15.0 12.0 27.0

Luxeon Flash6 LL60.0134 3222 023 63610 444 1,050,762 0.71 2.14 117,000 99% 746,041 20.0 15.0 35.0

Luxeon Flash6 LL60.0134 3222 023 63510 555 350,254 0.71 2.17 117,000 99% 248,680 7.0 5.0 12.0

Luxeon Rebel WW LL60.0187 3222 023 61200 98,129 1.36 3.43 90,000 99% 133,455 4.0 4.0 8.0

Luxeon Rebel Hikari 2700K 80CRI T&R LL60.0170 91,847 1.00 3.47 60,000 99% 91,847 6.0 4.0 10.0

Luxeon Rebel Hikari 3000K 80CRI T&R LL60.0174 3299 999 65100 147,404 1.00 3.56 60,000 99% 147,404 9.0 6.0 15.0

Luxeon Altilon C2 LL60-0156-018 3222 023 60100 1820 11,886 0.79 2.93 40,000 99% 9,390 1.0 1.0 2.0

Luxeon Altilon C2 - 3222 023 60400 1813 62,401 0.79 2.61 40,000 99% 49,297 5.0 3.0 8.0

Luxeon Altilon C2 - 3222 023 60300 1806 72,801 0.79 2.36 40,000 99% 57,513 5.0 4.0 9.0

Luxeon Rebel PC Amber LL60.125 3222 023 64200 159,863 0.84 6.00 95,000 99% 134,285 11.0 4.0 15.0

Total # Kanban 106 66 172

# Kanban tickets


Appendix L Glossary of terms

Term Description

APAC Asia Pacific Region

CLIP Confirmed Line Item Performance

CV Coefficient of variation (=standard deviation (σ) / average (μ))

CVP Confirmed Volume Performance

DC Distribution Centre

DP Demand Pattern

ERP Enterprise resource planning (an integrated software application)

GGI Gold to gold interconnect, production step at a Lumiled component supplier.

IDEF0 Integration DEFinition for Function

Kanban Also spelled kamban and literally meaning "signboard" or "billboard", is a concept

related to lean and just-in-time production.

LCD A liquid crystal display (LCD) is a thin, flat electronic visual display that uses the

light modulating properties of liquid crystals.

LCS Logistic department at the Component Supplier

LED A Light-Emitting Diode is a semiconductor light source.

LT Lead-time

MES Manufacturing Excellence System

MRP Master Requirement Plan

MPS Master Production Schedule

PP Production Plan

ROP Re-Order-Point

SADT Structured Analysis and Design Technique is a software engineering methodology

for describing systems as a hierarchy of functions.


SC Supply Chain

SS Safety Stock

S&OP Sales & Operations Planning

TIP Technology Institutions Process

WOW Way Of Working

WIP Work In Process

WW World Wide

Z Service Factor

Date post:	11-Sep-2021
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Supply Chain planning at Philips Lighting Lumileds

Documents