基于多目标遗传算法的维修保障资源调度优化研究  

Research on Maintenance Support Resources Optimization Based on Multi-objective Genetic Algorithm

作　　者：王涛[1,2] 张伏龙[2] 李广峰[2] 

机构地区：[1]北京理工大学机电学院,北京100081 [2]北京特种车辆研究所,北京100072

出　　处：《计算机测量与控制》2013年第11期3017-3020,共4页Computer Measurement &Control

摘　　要：以某型特种车辆为研究对象,其三级保养工艺流程优化属于多目标资源受限项目调度问题(Multi-Objective RCPSP,MORCPSP),根据车辆维修保养实际调度特点和决策者关心最大完工时间、人力资源总负荷和关键人力资源负荷的3个优化目标,设计了一种基于工序和人力资源的双重编码方案,并针对双重编码进行了遗传操作算子的改进,采用了NSGA-II算法精英保留策略和快速非支配排序构造Pareto最优解集,针对某型车辆三级保养工艺流程问题,算法可以有效求得其17个最优解,最后采用加权法对求得的Pareto最优解进行评定。This articles aims to research a type of special vehicle so as to optimize its technological process in ＂three--level＂ maintenance, which belongs to the issue of multi--objective resource constrained project scheduling problem （Multi -- Objective RCPSP, MORCPSP）. It puts forward a dual coding scheme based on processes and human resources, according to the actual scheduling of maintenance and three opti- mization objective, makespan, human resources overall load and the key human resource to load, that policymakers concerned. We improved the genetic operator and dual coding, adopted the elite reserved strategy and fast non--dominated sort of NSGA--II to construct the Pareto optimal solution set. The case application validates the correctness of the multi--objective algorithm to solve the maintenance support.

关 键 词：维修保障 多目标优化 遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]