上海交通大学学报

上海交通大学学报 ›› 2019, Vol. 53 ›› Issue (5): 600-609.doi: 10.16183/j.cnki.jsjtu.2019.05.013

• 学报（中文） • 上一篇    下一篇

考虑资源空窗期的资源投入问题的建模与优化

陆志强，石婷   

  1. 同济大学 机械与能源工程学院， 上海 201804
  • 出版日期:2019-05-28 发布日期:2019-05-28
  • 作者简介:陆志强（1968-），男，江苏省太仓市人，博士，教授，主要研究方向为物流与供应链建模与优化、生产工程等.
  • 基金资助:
    国家自然科学基金资助项目（61473211）

Modeling and Optimization of Resource Investment Problem with Resource Vacations

LU Zhiqiang,SHI Ting   

  1. School of Mechanical Engineering, Tongji University, Shanghai 201804, China
  • Online:2019-05-28 Published:2019-05-28

PDF (PC)

286

摘要： 针对飞机装配过程中部分关键资源存在不可用期的特征，研究基于资源空窗期的资源投入问题，建立以最小化资源使用成本为目标的作业调度数学模型.通过分析空窗期对作业开始时间决策区间的影响,设计以作业位置编码的遗传算法，充分利用迭代过程中得到的作业不同开始时间对应不同目标值的信息，提出基于概率分布的作业开始时间选择方法来改进变异操作，并通过部分作业执行顺序的分支枚举对所得结果进行局部优化.数据实验表明，对于小规模问题可获得近似精确解，而对于大规模问题比较现有的算法，其在算法求解精度上可提升3%.

关键词: 资源投入问题, 空窗期, 遗传算法, 分支定界算法

Abstract: Considering problem that some key resources are unavailable during aircraft assembly process, resource investment problem with resource vacations was studied. A mathematical model with objective function of minimizing the resource usage cost was built. By analyzing the influence of resource vacation on decision of job start time, a genetic algorithm was designed to code job positions. By making full use of the information in the iteration process that different start times corresponded to different target values, a method which determined the job start time based on the probability distribution was put forward to improve the mutation operation. And branch-and-bound algorithm was used to make local improvement. Comparative computational results reveal that the algorithm mentioned above can obtain approximate solution for small scale case, and improves the accuracy about 3% for the larger scale case compared with existing algorithms.

Key words: resource investment problem (RIP), resource vacation, genetic algorithm, branch-and-bound algorithm

中图分类号: