On optimal scheduling under uncertainty

Abstract

In this work we treat the problem of scheduling under two types of temporal uncertainty, set-based and probabilistic. For the former we define appropriate optimality criteria and develop an algorithm for finding optimal scheduling strategies using a backward reachability algorithm for timed automata. For probabilistic uncertainty we define and solve a special case of continuous-time Markov Decision Process. The results have been implemented and were applied to benchmarks to provide a preliminary assessment of the merits of each approach. © Springer-Verlag Berlin Heidelberg 2003.