计算机集成制造系统 ›› 2017, Vol. 23 ›› Issue (第3期): 567-574.DOI: 10.13196/j.cims.2017.03.014

• 产品创新开发技术 • 上一篇    下一篇

基于人工蜂群优化的串并行混装线关联排序问题

李修琳1,傅培华1,鲁建厦2,李进1   

  1. 1.浙江工商大学物流管理与工程系
    2.浙江工业大学工业工程研究所
  • 出版日期:2017-03-31 发布日期:2017-03-31
  • 基金资助:
    国家自然科学基金资助项目(71302035);浙江省自然科学基金资助项目(LQ14E050001,LZ14G020001);教育部人文社科资助项目(14YJA630046);浙江省教育厅资助项目(Y201330222)。

Integrated sequencing of serial /parallel mixed assembly line based artificial bee colony

  • Online:2017-03-31 Published:2017-03-31
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.71302035),the Zhejiang Provincial Natural Science Foundation,China(No.LQ14E050001,LZ14G020001),the MOE Project of Humanities and Social Sciences,China(No.14YJA630046),and the Zhejiang Provincial Education Department Project,China (No.Y201330222).

摘要: 为解决一类具有串并行混流装配结构的关联排序问题,建立了以串行线和总装线投产序列差异度最小、并行线和总装线投产序列差异度最小以及总装线物料消耗均衡为目标的多目标关联排序模型;在多目标问题的处理上,分别对串行线、并行线与总装线的序列差异根据序列调整成本设置了权重系数,同时设计了一种基于模糊目标规划的多目标人工蜂群优化策略对问题进行优化;建立了考虑串并行总差异度最小与总装线物料消耗均衡的隶属函数,设计了一种改进人工蜂群优化算法对隶属函数构建过程中的单目标优化问题以及转化后的集成优化问题进行了求解,其中针对多段编码设计了一种多段随机搜索的雇佣蜂寻优方法,提高了算法的全局寻优能力,引入适应性邻域,增强了算法的局部寻优能力;通过构造Benchmark算例验证了算法的有效性,并采用冰箱生产关联排序问题实例对模型与方法的有效性进行了验证。

关键词: 混流装配线, 关联排序, 人工蜂群优化算法, 模糊目标规划

Abstract: To solve the integrated sequencing problem of a kind of mixed-model assembly production with serial-parallel structure,a multi-objective sequencing model was established based on diversity factor minimization of serial line and diversity factor minimization of parallel line and level scheduling.To deal with the multi-objective problem,the weighting coefficients for sequence difference of serial line,parallel line and assembly line were set respectively according to the sequencing adjusting cost,and a multi-objective Artificial Bee Colony (ABC) optimization algorithm was proposed to optimize this problem based on Fuzzy Goal Programming (FGP).In this method,a membership function considering level scheduling and total diversity was established,and an improved ABC algorithm was proposed to solve the single goal problem and integrated problem after transformation in the process of function establishment.A multipart random search in employ-bee stage was put up to enhance the global optimization and an adaptive neighborhood was introduced to the algorithm to improve the local search performance.The validity of proposed algorithm was proved through an integrated sequencing instances constructed with Benchmark,and an instance simulation of freezer manufacturing was used to testify the validity of the model and method.

Key words: mixed assembly line, integrated sequencing, artificial bee colony algorithm, fuzzy goal programming

中图分类号: