A Physics Approach to Supply Chain Oscillations and Their Control

Abstract

Virtually all manufacturing involves supplies chains, in which value is added by each entity in the chain until a finished product emerges at the end. A serious problem that plagues all supply chains is unwanted fluctuation in the inventories along the chain. This disrupts the product output, is costly in capital, and can result in considerable disruption and hardship in personal lives. J.D. Sterman and his colleagues at MIT (Sterman & Fiddaman, 1993) have performed a useful service by providing business schools with a widely used simulation game that demonstrates for a beer distribution supply chain how easily the fluctuations can arise, and how difficult they can be to control. Their simulation results indicate that the oscillations are due to the over-reaction to input fluctuations by the individual entities in the chain, and the results strongly suggest that the situation could be improved by information technology that enables real-time feedback. The purpose of this chapter is to describe a statistical physics approach to understanding the supply chain oscillations. Standard techniques from statistical physics can lead to insights on the nature of the oscillations and on means for their control. Much of the supply chain work has been reported earlier, primarily at a number of conferences (Dozier a& Chang, 2004a, 2004b, 2005, 2006a, 2006b, 2007), but this chapter is the first time that the general approach together with its applications have been assembled in one place, along with a number of possible extensions. Section 2 provides some background on the application of statistical physics techniques to manufacturing issues. The focus in this section is on a pseudo-thermodynamics description of a manufacturing sector, and on the nature of effective intervention forces that can be uniquely derived from this thermodynamics. Section 3 demonstrates how the statistical physics approach leads directly to the existence of supply chain normal mode oscillations. It is shown that the nature of the normal modes depends on the extent of the information exchange between the entities in the supply chain. When the information exchange occurs only between an entity and the two entities immediately below and above it in the chain, the normal mode oscillation frequencies depend strongly on the way the oscillation amplitudes change along the chain: the normal modes resemble sound waves. On the other hand, when the information exchange occurs between an entity and all the other entities in the chain, the normal mode oscillation O pe n A cc es s D at ab as e w w w .in te ch w eb .o rg