复杂制造环境下的改进非支配排序遗传算法  

Improved NSGA for complex manufacturing environment

作　　者：刘爱军[1,2] 杨育[1] 程文明[2] 邢青松[1] 陆惠[3] 赵小华[1] 张煜东[4] 曾强[5] 姚豪 

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400030 [2]西南交通大学机械工程学院,四川成都610031 [3]上海师范大学天华学院,上海201815 [4]哥伦比亚大学脑图像实验室,纽约美国10032 [5]河南理工大学工业工程系,河南焦作454000 [6]马斯特里赫特大学知识工程系

出　　处：《计算机集成制造系统》2012年第11期2446-2458,共13页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(71071173);教育部高等学校博士学科点科研基金资助项目(20090191110004);新世纪优秀人才支持计划资助项目(NCET-07-0908);中央高校基本科研业务费科研专项资助项目(CDJZR10110012;SWJTU12BR018)~~

摘　　要：针对柔性作业车间多目标调度问题,在考虑机器、操作人员等资源约束和交货日期不确定性的基础上,构建了以加工成本、客户满意度及生产总流程时间为目标函数的模糊调度数学模型。针对传统的加权系数方法不能很好地解决柔性作业车间调度多目标优化问题的缺点,提出改进的非支配排序遗传算法,采用改进的拥挤密度排序法改善同一非劣等级内个体的排序;提出自适应交叉和变异策略,克服了种群早熟化,改善了算法的收敛速度;采用改进精英策略保持种群多样性,改善了算法的搜索性能。将该算法应用于某机械公司的人机双资源多目标柔性车间模糊调度,仿真结果证明了该方法的有效性和可行性。To solve the multi-objective optimization problem in flexible Job Shop scheduling, a fuzzy scheduling math- ematical model with objective functions of cost, total production cycle time and customer satisfaction was constructed by considering the resource constraints of machines and operators and the uncertainty of delivery date. Aiming at the problem that traditional weighted coefficient method could not solve the multi-objective scheduling optimization, an improved Non-dominated Sorting Genetic Algorithm（NSGA-I] ）was proposed. In this algorithm, an improved crow- ding density sorting method was used to ameliorate the individual sorting in same non-inferior grade; an adaptive crossover and mutation strategy was proposed to overcome the prematurity of population and to improve convergence speed. Improved elitism strategy was adopted to ensure the population diversity and improve the search perform-anee. The algorithm was applied in a man-machine dual resource and multi-objective flexible workshop fuzzy schedu- ling in a machine company, and the feasibility and efficiency of algorithm were verified.

关 键 词：柔性车间调度 多目标优化 改进非支配排序遗传算法 仿真 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置]