AN EVALUATION OF SIMULATED ANNEALING FOR BATCH PROCESS SCHEDULING

In recent years, simulated annealing has been successfully used to solve several combinatorial optimization problems. Most scheduling problems in batch processing are also combinatorially complex. In this paper, we investigate the usefulness of simulated annealing for solving batch process scheduling problems. For this, we develop a simulated annealing methodology for minimizing the total time to produce a set of batches in the serial flowshop with unlimited storage. We compare the method with the best heuristic method and two other random strategies. Its generality, simplicity and near-optimal nature of its solutions far outweigh its large computational requirements. In fact, it seems hard to devise a heuristic algorithm that can give solutions as good as those given by the simulated annealing, even if we allow more computational effort. Simulated annealing appears to be a very versatile and powerful method for solving different forms of the batch process scheduling problem.