This book summarizes a set
of lectures given at U.C. Berkeley in the Fall of 2001, highlighting the connection
between traffic flow, queuing systems and supply chains. The book unveils the root causes
of the "bullwhip" effect; i.e., where the production of raw materials in a
supply chain is seen to be more volatile than that of intermediate goods, and even more so
than the final customer demand. It also shows that this indesirable effect arises if
suppliers act in a certain non-cooperative way, even if they have perfect information
about the future. Their actions, just as importantly as what they know, determine the
stability of a supply chain. The book then describes control methods for eliminating all
instabilities without increasing supplier costs, and presents approximate cost formulas.
It also shows that for every supply chain, there is a dual queuing system with identical
behavior, and that queuing systems can be similary controled.
Contents
Introduction
Algorithms/Policies
Algorithmic Properties
Stability and Monotonicity
Requirements
Strongly Stable Policies:
The Act Method
Cost Estimation and
Optimization
Discussion
References
Appendix A: Stability via
Control Theory
Appendix B: Kinematic Wave
Theory Revisited
123 pages