摘要
探讨了双目标下,带一种资源约束的,工件成类别的并行机器调度问题.针对该问题,提出了一种遗传算法.该算法采用了两两竞赛的选择算子、聚集度、违约度来处理多目标约束优化.通过随机订单的测试,计算结果显示:对于各个单目标值,该算法比修正的EDD、LPT、SPT能改善3%～37%.
Multi-objective scheduling on parallel machines with families and one resource constrained is considered. The objectives are total tardiness and maximum make-span. This paper proposes genetic algorithms using tournament selection, niche count, and violation degree to solve this problem. The computational results from random orders show that the genetic algorithms are more efficient than rectified EDD, LPT, and SPT in solving the problem.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2005年第9期78-82,共5页
Systems Engineering-Theory & Practice
关键词
多目标
约束
并行机器
遗传算法
multi-objective
constraint
parallel machines
genetic algorithms
