基于改进多目标微粒群算法的模具多项目反应调度

Reactive Scheduling for Multiple Mould and Die Projects Based on Improved Multi-objective Particle Swarm Optimization

针对模具多项目执行过程中由于可再生资源发生故障、任务拖期以及随机插入新项目而导致的调度计划变更问题,提出了一种反应调度算法。利用生灭过程理论对可再生资源的不确定性进行分析,并采用基于优先规则的拓扑排序方法构建初始调度计划,建立了以调度计划变更产生的附加费用最小以及项目加权工期之和最小为目标的多目标反应调度模型,并提出一种改进的多目标微粒群算法进行模型求解。最后,通过仿真计算分析了反应调度算法的可行性与有效性。

This paper proposed a reactive scheduling algorithm for multiple mould and die projects to deal with multiple stochastic disruptions during projects execution including uncertain renewable resources infeasibilities due to breakdowns,multiple tasks taking longer time than planning and random arrivals of new projects.Firstly,uncertain renewable resources availabilities were analyzed with the theory of birthdead process,and an initial schedule was constructed by using topological sorting method based on priority rules.Then,a multi-objective reactive scheduling model with the optimization objects of minimizing the weighted sum duration of projects and the disruptions cost,defined as the weighted sum of the deviations between the executing schedule before reactive scheduling and a new schedule after reactive scheduling was built,and an improved multi-objective particle swarm optimization was proposed to solve it.Finally,the feasibility and reliability of the reactive scheduling algorithm were analyzed by simulations.