引用本文:彭涛,周炳海.车辆装配线线边物料准时化配送算法[J].控制理论与应用,2016,33(6):779~786.[点击复制]
PENG Tao,ZHOU Bing-hai.Just-in-time distribution algorithm of line-side parts for automobile assembly lines[J].Control Theory and Technology,2016,33(6):779~786.[点击复制]
车辆装配线线边物料准时化配送算法
Just-in-time distribution algorithm of line-side parts for automobile assembly lines
摘要点击 3199  全文点击 2066  投稿时间:2015-07-30  修订日期:2016-03-17
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2016.50657
  2016,33(6):779-786
中文关键词  准时化  调度  启发式算法  疫苗接种  Metropolis准则
英文关键词  just-in-time  scheduling  heuristic algorithms  vaccine inoculation  Metropolis criteria
基金项目  国家自然科学基金(61273035;71471135)
作者单位E-mail
彭涛 同济大学 tao.100824@foxmail.com 
周炳海* 同济大学 bhzhou@tongji.edu.cn 
中文摘要
      为了有效地解决带线边缓存容量约束的车辆装配线准时化物料配送问题, 提出了改进免疫克隆选择调度 算法. 首先进行了问题域的描述, 以最小化搬运成本和线边库存成本总和为优化目标, 建立了数学规划模型. 针对 这一多重约束的混合优化问题, 给出了问题的性质分析, 并将原问题转化为带缓存容量约束的离散优化问题. 在算 法设计过程中, 开发了直接反映配送路径及配送量的变长双层整数编码机制. 为了提升算法的收敛性能, 在邻域生 成机制中融入了疫苗接种操作和基于Metropolis接受准则的局部搜索算子. 最后, 对算法进行了仿真实验. 结果表 明, 该配送算法可行、有效.
英文摘要
      Given the limited capacities of the line-side buffers, a modified immune clonal selection algorithm was proposed to investigate the just-in-time (JIT) part distribution for automobile assembly lines. Firstly, a JIT distribution problem domain was described. And the mathematical programming model was established with an objective of minimizing the sum of the inventory holding cost and delivery cost. For this composite optimization problem with multiple constraints, the relevant properties were analyzed to transform the original problem to the discrete optimization model with buffer capacities constraints. A two-level integer encoding mechanism with variant length was put forward in the algorithm design process, which reflected directly distribution routing and quantities.To improve the convergence performance of the proposed algorithm, the vaccine inoculation operator, as well as the local search schema based on Metropolis acceptance criterion, was introduced into the neighbourhoods generation mechanism. Finally, simulation experiments of the algorithm were carried out and the results demonstrate that the proposed method is feasible and effective.