基于改进遗传算法的网状流水线作业计划

Scheduling for Crossed Flow Shop Based on Modified Genetic Algorithm

在分析实际生产中流水作业种类的基础上提出了网状流水线作业计划的概念,它由两条或两条以上多阶段平行流水线构成;相邻阶段若干流水线之间存在交叉,在该交叉处前阶段加工完成的工件可向多条流水线后续设备流动;在传统遗传算法中引入多阶段编码、虚基因、自适应交叉操作等方法构成了改进遗传算法,以适应网状流水线作业计划的需要;建立了基于改进遗传算法的网状流水线作业计划方法,使各阶段网状流水线之间工件数动态平衡;用算例证明了网状流水线的优点和算法的有效性。

On the basis of analysing the practical flow shops, the concept of crossed flow shop (CFS) was put forward. CFS was constituted of multiple stage parallel flow lines. There were intersections among adjacent flow lines. Jobs finished in the flow lines before intersections had the chances to move to one of the following flow lines. Modified genetic algorithm (MGA) for scheduling of CFS was deduced, with multi-stage encoding, virtual gene and adaptive crossover being introduced to traditional GA, which will achieve the dynamic balance of work-piece number in different production lines of the CFS. At the end, an experimental optimization verifies the advantages of CFS and the effectiveness of MGA.