Modeling and Optimization of Production Processes: Lessons from Traffic Dynamics

We will develop and study models of supply networks and how they relate to vehicular traffic. These models allow us to take into account the nonlinear, dynamical interactions of different production units and to test alternative management strategies with respect to their potential impacts. In this way, one can understand the preconditions of the so-called bull-whip effect (i.e., the fact that small variations in the consumption rate can cause large variations in the production rate of companies generating the requested product). Moreover, we will show how the nonlinear dynamics of a particular supply chain in semiconductor production has been optimized by means of the “slower-is-faster effect”' known from panicking pedestrian crowds. Driven many-particle models of pedestrian motion also offer solutions for other typical problems of nonlinear production processes such as the coordination of robots, the efficient segregation of different kinds of objects, or the frictionless merging of object flows at bottlenecks. Finally, from the simulation of pedestrian behavior one can learn how fluctuations could be used to increase the order in the system, how to speed up certain production processes, or how to compensate for delays in a series of production steps.