Parallelism for perturbation management and robust plans

Abstract

An important insufficiency of modern industrial plans is their lack of robustness. Disruptions prevent companies from operating as planned before and induce high costs for trouble shooting. The main reason for the severe impact of disruptions stems from the fact that planners do traditionally consider the precise input to be available at planning time. The Repair Game is a formalization of a planning task, and playing it performs disruption management and generates robust plans with the help of game tree search. Technically, at each node of a search tree, a traditional optimization problem is solved such that large parts of the computation time are blocked by sequential computations. Nevertheless, there is enough node parallelism which we can make use of, in order to bring the running times onto a real-time level, and in order to increase the solution quality per minute significantly. Thus, we are able to present a planning application at the cutting-edge of Operations Research, heavily taking advantage of parallel game tree search. We present simulation experiments which show the benefits of the repair game, as well as speedup results. © Springer-Verlag Berlin Heidelberg 2005.