摘要
针对单机Just-in-time系统在加工时间有较大范围不确定性的调度环境,设计出一种两层协同进化遗传算法,解决绝对鲁棒调度优化问题,最优化在加工时间变化范围内预调度的最差性能.外层遗传算法确定工件的加工顺序,内层遗传算法确定在给定调度顺序下实现最差性能的加工时间.通过对大量随机算例进行仿真,并与采用期望加工时间的确定性调度算法进行对比,表明所提出的算法是有效的.
A two-loop co-evolutionary genetic algorithm is proposed to solve the absolute robust scheduling for a Just-in-time single machine with significant processing time uncertainty. The worst-case performance of a predictive schedule over the range of job processing times is optimized. The outer loop of the proposed algorithm is to determine the job sequence on machine and the inner loop searches for the processing time scenario with worst-case performance for a given sequence. The simulation results show the proposed method is very effective compared with the deterministic scheduling method based on expected job processing times.
出处
《控制与决策》
EI
CSCD
北大核心
2007年第10期1151-1154,1159,共5页
Control and Decision
基金
国家自然科学基金项目(60504026)
上海市科技发展基金项目(04DZ11008)
关键词
遗传算法
鲁棒调度
最差性能
Genetic algorithm
Robust scheduling
Worst-case performance
