作　　者： ; ; ; ;

机构地区： 广东工业大学机电工程学院

出　　处： 《系统工程理论与实践》 2011年第8期1571-1580,共10页

摘　　要： 针对模具多项目执行过程中由于任务拖期而导致的调度计划变更,提出了一种基于优化的资源流约束的反应调度算法.首先利用基于优先规则的微粒群算法构建一个项目加权工期之和最小的初始调度计划,采用启发式算法建立并优化初始调度计划的资源流网络,并将关键链技术与资源流网络相结合,对初始调度计划进行合理的时间缓冲设置.然后建立了以调度计划扰动费用最小为优化目标的反应调度模型,并用基于优先规则的微粒群算法进行求解.最后,通过仿真计算分析了算法的可行性与可靠性,结果表明该反应调度算法在模具多项目动态调度中具有一定的应用价值. This paper proposed a reactive scheduling algorithm based on optimized resource flow constraints, which was used to repair multiple mould and die projects baseline schedule that suffer from multiple tasks taking long time than planning during projects execution.Firstly,a baseline schedule minimizing weighted sum duration of projects was built with priority rules based particle swarm optimization and a resource flow network was built and optimized for the baseline schedule with heuristic algorithm. By combining critical chain technology and resource flow network,the baseline schedule was improved by setting proper time buffers.Then,a heuristic reactive scheduling model with the optimization object of minimizing the disruptions cost when tasks starting time changed during projects execution was constructed and priority rules based particle swarm optimization was used to solve it.Finally,the feasibility and reliability of the above reactive algorithm were analyzed by simulations and the results show that this algorithm could effectively solve dynamic scheduling problems for multiple mould and die projects with multiple tasks duration enlarging.

关 键 词： 资源流 模具多项目调度 反应调度 关键链 微粒群算法 混沌序列

领　　域： [机械工程]