Authors: Haipeng Zhang, Mitsuo Gen
Addresses: Department of Industrial Engineering, Pusan National University, Korea. ' Graduate School of Information, Production and Systems, Waseda University, Japan
Abstract: In this paper, we combine ACO with some randomised dispatching heuristics and propose a special transition rule for finding the best schedule to the JSP problems. Moreover, a special critical path-based local search is also combined to improve the best solutions by reducing the idle time. In order to gain higher efficiency of the proposed algorithm and avoid the early convergence in local optimal solution, we enhance the hybrid ACO by building a parallel hybrid Ant Colony Optimisation (ph-ACO) algorithm. Some numerical examples are used to demonstrate the performance of the ph-ACO and we can find that the proposed ph-ACO algorithm with Longest Remaining processing Time (LRT) and Longest Remaining processing time Excluding the operation under consideration (LRE) can both improve the efficiency of the algorithm obviously. Furthermore, we also decide the appropriate parameter setting of β is around 2. Finally, after comparing with hybrid Genetic Algorithm (GA) by solving same benchmark problems, the experimental results show the proposed ph-ACO outperforms traditional ACO and hybrid GA.
Keywords: ant colony optimisation; ACO; metaheuristics; genetic algorithms; GAs; critical path; dispatching heuristics; job shop scheduling.
International Journal of Manufacturing Technology and Management, 2009 Vol.16 No.1/2, pp.22 - 41
Published online: 30 Nov 2008 *Full-text access for editors Access for subscribers Purchase this article Comment on this article