摘要
目的解决单一粒子群算法求解Job Shop调度问题存在的不足,提高这类问题的求解质量.方法采用粒子群算法进行全局搜索,将禁忌搜索算法用于并行局部搜索,禁忌搜索在找到改进解的邻域时采用动态记忆的方式.结果在较短时间内,找到了LA21,LA24等典型benchmarks问题的最优解,十次求解的平均值的平均相对误差百分比比并行遗传算法和禁忌搜索算法分别小了2.94%和0.56%.结论提出一种混合粒子群算法,增强了粒子群算法的局部搜索能力,说明该混合粒子群优化算法是有效的.
The key of the paper is to offset the deficiency in the solution of Job Shop scheduling and to improve the quality of the solution of such problems. We applied Particle Swarm Optimization （PSO） algorithm to global search and taboo search algorithm to local search. The taboo search algorithm stores current solution dynamically after searching the neighborhood of improved solution. The best solutions of typical benchmark problems such as LA21, LA24 were searched in shorter time. And the average relative error percentage of the ten- time average value is respectively 2.94 % and 0.56 % smaller than that of the Parallel Genetic Algorithm and Taboo Search Algorithm. A hybrid PSO Algorithm is proposed, which has intensified the local search ability of PSO and we can conclude that hybrid particle swarm optimization algorithm is ef- fective.
出处
《沈阳建筑大学学报（自然科学版）》
EI
CAS
2008年第3期494-498,共5页
Journal of Shenyang Jianzhu University：Natural Science
基金
辽宁省教育厅资助项目(20060701)
