计算机集成制造系统 ›› 2014, Vol. 20 ›› Issue (6): 1398-1404.DOI: 10.13196/j.cims.2014.06.cuiweiwei.1398.7.20140617

• 产品创新开发技术 • 上一篇    下一篇

基于多目标优化的生产调度与设备维护集成研究

崔维伟1,陆志强2+,潘尔顺1   

  1. 1.上海交通大学工业工程与物流管理系
    2.同济大学机械学院
  • 出版日期:2014-06-30 发布日期:2014-06-30
  • 基金资助:
    国家自然科学基金资助项目(71171130);上海市自然科学基金资助项目(12ZR1414400)。

Production scheduling and preventive maintenance integration based on multi-objective optimization

  • Online:2014-06-30 Published:2014-06-30
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.71171130),and the Natural Science Foundation of Shanghai Municipality,China(No.12ZR1414400).

摘要: 为了解决生产实际中工件调度与维修计划的相互影响问题,提出基于多目标遗传算法的联合优化方案,以单机系统为研究对象,设备失效函数服从威布尔分布,考虑机器和工件的堕化效应,综合决策工件加工顺序和预防性维护时间。以工件流程时间最短化和维修成本最小化为联合优化目标,基于非支配排序遗传算法框架,提出一种新的选择机制以及去除重复个体的方法以提高种群多样性,设计改进的多目标遗传算法以求解Pareto最优解。通过不同设置下的数据实验验证了基于多目标优化的联合决策比独立决策表现更优异。实现了生产与维修部双目标之间的权衡,使决策者可根据偏好选择不同的满意解,有效协调车间的生产调度与设备维护计划。

关键词: 调度, 预防性维护, 多目标优化

Abstract: To solve the interrelationship between the production scheduling and maintenance planning,a joint method based on multi-objective genetic algorithm was proposed.Aiming at the single machine system,a failure function governed by Weibull was studied.By considering the deterioration effect of jobs and machine,the processing sequence of jobs and preventive maintenance times of machine were determined comprehensively.To minimize the makespan and maintenance cost simultaneously,a modified multi-objective genetic algorithm based on NSGA-II was proposed to optimize the Pareto front,in which a new selection method and an eliminating overlapping solutions method were embedded to increase the population diversity.The computational experiments under different problem setting showed that the joint decision making based on multi-objective optimization was better than the independent decision making.The double objectives of production and maintenance departments were well balanced,and the decision maker could coordinate the production scheduling and maintenance planning efficiently according to the different bias solutions.

Key words: scheduling, preventive maintenance, multi-objective

中图分类号: