Proactive algorithms for scheduling with probabilistic durations

Abstract

Proactive scheduling seeks to generate high quality solutions despite execution time uncertainty. Building on work in [Beck and Wilson, 2004], we conduct an empirical study of a number of algorithms for the job shop scheduling problem with probabilistic durations. The main contributions of this paper are: the introduction and empirical analysis of a novel constraint-based search technique that can be applied beyond probabilistic scheduling problems, the introduction and empirical analysis of a number of deterministic filtering algorithms for probabilistic job shop scheduling, and the identification of a number of problem characteristics that contribute to algorithm performance.