针对作业调度问题的遗传选择算子性能分析

Analysis of genetic selection operator performance in job shop scheduling problem

针对作业调度问题,通过变形遗传算法实验,对轮盘赌、随机联赛、随机遍历抽样和确定式采样等选择算子进行了比较分析。以FT06典型车间作业调度问题为实例,比较了这几种常用选择算子在解决车间作业调度问题时的性能优劣程度;从全局收敛性和收敛速度两个方面,分析总结了这些选择算子对算法的全局搜索能力的影响程度。实验结果表明随机遍历抽样算子的整体性能要优于其他几种选择算子。

For job shop scheduling problem,a comparative analysis was made on genetic selection operator including roulette wheel selection,stochastic tournament model,stochastic universal sampling,deterministic sampling by experiments of deformative genetic algorithm.Using FT06 of typical example on job shop scheduling problem,these common selection operators were compared in solving job shop scheduling problem to evaluate the quality of performance.From the global convergence and convergence rate,the effect of these selection operators on the global search capability was summarized.The results show that the overall performance of stochastic universal sampling operator was better than other selection operators.