A practical decomposition approach for a chemical substance scheduling problem

Abstract This paper describes a decomposed iterative improvement scheduling algorithm developed to solve a scheduling problem of a chemical substance production machine. The algorithm has to determine the most effective schedule to produce several chemical substances. In order to build optimal feasible schedules a hybridization of AI techniques (heuristics, simulated annealing, constraint propagation) is used in a practical algorithm architecture. The algorithm architecture is compared against human schedule generation and a simple opportunistic algorithm.