Sethi / Yan / Zhang Inventory and Supply Chain Management with Forecast Updates

1. Auflage 2006
ISBN: 978-0-387-25663-4
Verlag: Springer US
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, Band 81, 292 Seiten

Reihe: International Series in Operations Research & Management Science

Real problems are formulated into tractable mathematical models, which allow for an analysis of various approaches. Attention is focused on solutions. Provides a unified treatment of the models discussed , presents a critique of the existing results, and points out potential research directions.

Autoren/Hrsg.

Sethi, Suresh P.

Yan, Houmin

Zhang, Hanqin

Inhaltsverzeichnis

1;Contents;6
2;List of Figures;11
3;List of Tables;12
4;Preface;14
5;Notation;18
6;Chapter 1 INVENTORY AND SUPPLY CHAIN MODELS WITH FORECAST UPDATES;20
6.1;1.1. Introduction;20
6.2;1.2. Aims of the Book;21
6.3;1.3. Information Dynamics in Supply Chains;24
6.4;1.4. Inventory and Supply Chains with Multiple Delivery Modes;29
6.5;1.5. Supply Contracts;31
6.6;1.6. Competitive Supply Chains;33
6.7;References;36
7;Chapter 2 EXAMPLES FROM INDUSTRY;42
7.1;2.1. Introduction;42
7.2;2.2. Industry Observations;43
7.3;2.3. Multistage Forecasts;50
7.4;2.4. Operational Factors Affecting Forecasting Process;52
7.5;2.5. Concluding Remarks;60
7.6;2.6. Notes;61
8;Chapter 3 INVENTORY MODELS WITH TWO CONSECUTIVE DELIVERY MODES;64
8.1;3. I. Introduction;64
8.2;3.2. Notation and Model Formulation;65
8.3;3.3. Dynamic Programming and Optimal Nonanticipative Policy;70
8.4;3.4. Optimality of Base- Stock Policies;78
8.5;3.5. The Nonstationary Infinite- Horizon Problem;88
8.6;3.6. An Example;95
8.7;3.7. Concluding Remarks;103
8.8;3.8. Notes;103
9;Chapter 4 INVENTORY MODELS WITH TWO CONSECUTIVE DELIVERY MODES AND FIXED COST;108
9.1;4.1. Introduction;108
9.2;4.2. Notation and Model Formulation;109
9.3;4.3. Dynamic Programming and Optimal Nonanticipative Policy;111
9.4;4.4. Optimality of (s, S) Ordering Policies;113
9.5;4.5. Monotonicity Properties;134
9.6;4.6. The Nonstationary Infinite- Horizon Problem;140
9.7;4.7. Concluding Remarks;143
9.8;4.8. Notes;144
9.9;References;145
10;Chapter 5 INVENTORY MODELS WITH THREE CONSECUTIVE DELIVERY MODES;148
10.1;5.1. Introduction;148
10.2;5.2. Notation and Model Formulation;149
10.3;5.3. Dynamic Programming and Optimal Nonanticipative Policies;155
10.4;5.4. Optimality of Base- Stock lype Policies;163
10.5;5.5. The Nonstationary Infinite-Horizon Problem;177
10.6;5.6. Concluding Remarks;180
10.7;5.7. Notes;181
10.8;References;182
11;Chapter 6 MULTIPERIOD QUANTITY- FLEXIBILITY CONTRACTS;184
11.1;6.1. Introduction;184
11.2;6.2. Model and Problem Formulation;185
11.3;6.3. Contingent Order Quantity at Stage 2;189
11.4;6.4. Optimal Purchase Quantity at Stage 1;194
11.5;6.6. Multiperiod Problems;221
11.6;6.7. Numerical Example;225
11.7;6.8. Concluding Remarks;235
11.8;6.9. Notes;237
12;Chapter 7 PURCHASE CONTRACT MANAGEMENT: FIXED EXERCISE COST;242
12.1;7.1. Introduction;242
12.2;7.2. Problem Formulation;243
12.3;7.3. Optimal Solution for Stage 2;245
12.4;7.4. Optimal Solution for a Class of Demand Distributions;251
12.5;7.5. Analysis for Uniformly Distributed Demand;255
12.6;7.6. Concluding Remarks;272
12.7;7.7. Notes;273
13;Chapter 8 PURCHASE CONTRACT MANAGEMENT: TWO- PLAYER GAMES;276
13.1;8.1. Introduction;276
13.2;8.2. Problem Formulation;277
13.3;8.3. Reaction Strategies Under Uniformly Distributed Demand;281
13.4;8.4. A Static Noncooperative Game;286
13.5;8.5. A Dynamic Noncooperative Game;295
13.6;8.6. Concluding Remarks;300
13.7;8.7. Notes;302
14;Copyright Permissions;304
15;Index;306

Leseproben

2.2. Industry Observations (p. 24-25)
A major security-system manufacturing company produces and distributes security systems for military, residential, commercial, and industrial applications. It has a design center in California, a manufacturing center in Asia, and three regional distribution centers in San Francisco, Amsterdam, and Singapore. The company sources components and subassemblies around the world. The management objectives are to improve the response time to meet market demand, to reduce inventory, and to shorten lead time (including the time for manufacturing and distribution). In the security-system market, customers expect to have the required device or system within one month. Therefore, given long lead times in procurement and production, the manufacturing operation relies largely on forecasts.

From a practical point of view, forecasts are never accurate, and the company updates its demand forecasts until the real demand is realized. When too little raw material is ordered, the company has to pay a higher price to secure them or use air shipment to expedite them (if these options are feasible). When too many raw materials and subassemblies are ordered, the company has to keep them in inventory. These materials often become obsolete. These updates in forecasting also make it difficult for the company to allocate its production capacity efficiently.

A key component in security systems is the microcontroller, which makes up 30% to 40% of the total materials cost. A microcontroller is a central processing unit (CPU) chip with a built-in memory and interface circuits. The read-only memory (ROM) contains permanent data (program code). See Spasov [5] for a discussion of related concepts about microcontrollers and their technology. The company can order microcontrollers with user-supplied data requirements. If user-supplied data is provided, the semiconductor manufacturing includes a process known as custom photo masking in the wafer-fabrication process. Alternatively, the company can purchase microcontrollers with a programmable ROM such as one-time-programmable (OTP) read-only memory or erasable programmable read-only memory (EPROM). The company inputs the data into these programmable microcontrollers after the chips are received. To order custom-masked chips, the users are required to provide the data (program code) prior to manufacturing, and a significant lead time is required. On the other hand, since programmable ROMs are generic, these microcontrollers can be produced with a considerably shorter lead time. However, the OTP chips are about twice as expensive as custom-masked chips and EPROM chips are even more expensive. The company must decide how to order both custom-masked and OTP chips.

The company uses a half-year rolling window for demand forecasting. These forecasts are made and updated monthly by the regional offices. The headquarter coordinates the forecasts and passes them to its logistics and manufacturing functions. Procurement decisions are made based on the demand forecast and the lead time required by its vendors. The company divides the raw materials into two classes: critical and regular. The components that have fewer sources, and have a higher value content, and require a longer lead time are classified into their critical materials. Microcontrollers are a typical example.