• 论文 •    

一种求解变速机调度问题的混合蚁群优化算法

周泓,李政道,吴学静   

  1. 北京航空航天大学 经济管理学院,北京100083
  • 出版日期:2008-09-15 发布日期:2008-09-25

Hybrid ant colony optimization algorithm for unrelated parallel machine scheduling problem

ZHOU Hong,LI Zheng-dao,WU Xue-jing   

  1. School of Economics & Management, Beihang University, Beijing 100083, China
  • Online:2008-09-15 Published:2008-09-25

摘要: 针对一类变速机总加权拖期调度问题,提出一种混合蚁群优化算法。引入单机拖期调度问题中性能良好的修正预计完成时间的一种修改版本启发式规则,计算信息素初值,有利于算法跳出局部极值,并在局部搜索阶段,采用单亲遗传算法基因移位算子,有效优化当代最优解。通过均匀试验设计和统计分析,确定算法的关键参数组合,将算法应用于随机生成的不同规模的40个算例,并将其结果与同类文献中算法的优化结果进行对比分析。结果表明,在相同迭代次数下,混合算法优于对比算法。

关键词: 变速机, 总加权拖期, 调度, 启发式规则, 蚁群优化, 单亲遗传算法

Abstract: Aiming at the total weighted tardiness scheduling problem of unrelated parallel machine, a kind of hybrid ant colony optimization algorithm was proposed. Variation of Modified Due Date ( VMDD ) heuristic rules were introduced for pheromone initialization to help escape from local optimal, and Partheno-Genetic Algorithms ( PGA ) gene transfer operator was incorporated into local search to optimize the currently found best solution. Key parameters of the algorithm were defined by the uniform experimental design & statistic analysis. The algorithm was applied in 40 stochastic generation problems with different scale, and results were compared. Results demonstrated that the hybrid algorithms significantly outperformed the competitor with the same iteration.

Key words: unrelated parallel machine, total weighted tardiness, scheduling, heuristic rule, ant colony optimization, partheno genetic algorithm

中图分类号: