• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2020, Vol. 56 ›› Issue (14): 245-253.doi: 10.3901/JME.2020.14.245

• 交叉与前沿 • 上一篇    下一篇

扫码分享

基于两种改进差分进化的可修备件多级库存优化算法研究

顾涛, 李苏建   

  1. 北京科技大学机械工程学院 北京 100083
  • 收稿日期:2020-01-11 修回日期:2020-06-05 出版日期:2020-07-20 发布日期:2020-08-12
  • 作者简介:顾涛,男,1985年出生,博士研究生。主要研究方向为供应链计划优化、备件库存优化、生产调度优化等。E-mail:babygo1003@163.com;李苏建,男,1959年出生,博士,教授,博士研究生导师。主要研究方向为物流信息系统、企业物流管理等。E-mail:lisujian@me.ustb.edu.cn
  • 基金资助:
    国家部委资助项目(JCKY2018209C002)。

Research on Multi-level Inventory Optimization Algorithm of Repairable Spare Parts Based on Two Improved Differential Evolution

GU Tao, LI Sujian   

  1. School of Mechanical Engineering, University of Science and Technology Beijing, Beijing 100083
  • Received:2020-01-11 Revised:2020-06-05 Online:2020-07-20 Published:2020-08-12

摘要: 针对传统边际分析法求解多级可修备件库存模型解质量不高的问题,提出两种改进差分进化算法对模型进行求解,一种是带局部搜索的改进差分进化算法,另一种是基于边际分析法的改进差分进化算法。两种算法分别运行了20次,每次迭代上限设置为5 000次,得到相同的最优解,该解与已发表文献采用边际分析法求出的最优解相比库存总经费降低了4.44%,说明了两种算法具有一定的优越性。另外,基于边际分析法的改进差分进化算法较带局部搜索的改进差分进化算法具有明显的优越性,其中库存总经费均值低2.4%、库存总经费标准差低63.8%、迭代次数均值少38.7%,说明基于边际分析法的改进差分进化算法在优化水平、算法稳定性以及算法计算效率三个方面优于带局部搜索的改进差分进化算法。

关键词: 可修备件, 多级库存, 边际分析法, 改进差分进化算法

Abstract: In view of the low quality of the traditional marginal analysis method to solve the multi-level repairable spare parts inventory model, two improved differential evolution algorithms are proposed to solve the model, one is the improved differential evolution algorithm with local search, the other is the improved differential evolution algorithm based on the marginal analysis method. The two algorithms run 20 times respectively, and the upper limit of each iteration is set to 5 000 times, and the same optimal solution is obtained. Compared with the optimal solution obtained by marginal analysis method in published literature, the total inventory cost is reduced by 4.44%, which shows that the two algorithms have certain advantages. In addition, the improved differential evolution algorithm based on the marginal analysis method has obvious advantages over the improved differential evolution algorithm with local search, among which the average value of total inventory cost is 2.41%, the standard deviation of total inventory cost is 63.8% and the average number of iterations is 38.7%. It shows that the improved differential evolution algorithm based on the marginal analysis method has three advantages:optimization level, algorithm stability and algorithm calculation efficiency. It is better than the improved differential evolution algorithm with local search.

Key words: repairable spare parts, multilevel inventory, marginal analysis method, improved differential evolution algorithms

中图分类号: