Parallel simulated annealing for scheduling agile manufacturing systems

The problem addressed in this research is the scheduling of an agile manufacturing system comprised of machining and assembly operations with the objective of minimizing the makespan. This problem has been addressed in the literature using heuristics and metaheuristics. In addition, a lower bound has been developed. This research builds upon the findings of the recent literature and proposes two enhanced algorithms, parallel simulated annealing and parallel modified simulated annealing, to solve the addressed problem. The performance of the developed algorithms is compared to that of another simulated annealing algorithm that is available in the literature. Performance comparisons are based on the deviation above the lower bound and the frequency of achieving the best solution amongst all algorithms. The statistical paired t-test and a 2 cubed designed factorial experiment are utilized in these comparisons. The results demonstrate that the two algorithms developed outperform the existing SA algorithms. Factors affecting the performance of all algorithms are also investigated