工业工程

工业工程 ›› 2015, Vol. 18 ›› Issue (4): 140-145.

• 实践与应用 • 上一篇    下一篇

考虑均衡负载的车辆路径问题及算法设计

  

  1. （西南交通大学 交通运输与物流学院，四川 成都 610031）
  • 出版日期:2015-08-30 发布日期:2015-10-15
  • 作者简介:贺政纲（1977-），男，湖南省人，副教授，博士，主要研究方向为逆向物流、物流系统规划.
  • 基金资助:

    国家社会科学基金资助项目（13CGL127）；四川省科技厅软科学研究计划资助项目（2013ZR0041）；四川省哲社资助项目（SC13C042）

A Research on the Loadbalanced Vehicle Routing Problem and Algorithm Design

  1. (School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 610031, China)
  • Online:2015-08-30 Published:2015-10-15

摘要： 针对目前物流配送系统中存在的车辆负载不均衡导致的物流配送质量和配送系统柔性下降这一现象，提出了考虑均衡车辆负载的多目标路径优化模型，以配送车辆总行驶距离尽可能短和车辆之间载运量尽可能平衡为优化目标。〖JP2〗针对模型设计遗传算法，通过在变异操作后加入进化逆转操作，加强局部寻优，找到全局最优解。以1个配送中心和14个客户的配送网路为例，利用Matlab软件建模分析，设置并调节不平衡处罚参数，得到总行驶距离和不平衡度之间的变化关系图和若干备选决策。最后构建评价体系对各个决策进行评分，〖JP〗得出当总行驶距离为40819 5，不平衡度为1时，评分最高，为本算例最优解，表明该模型在实际应用中能够为管理者提供决策依据。

关键词: 车辆路径问题(VRP), 均衡负载, 遗传算法, 不平衡度

Abstract: With the unbalanced loads among vehicles causing the reduction of distribution quality and flexibility in the logistics activities, a multi-objective model is built, whose goal is to simultaneously minimize the total travel distance and imbalance penalty when load balancing among all vehicles is considered. The genetic algorithm mended by adding revolution operation after mutation operation is applied to solve the problem, which can avoid the algorithm dropping into local optimum and help the global search. A distribution system composed of one distribution center and 14 customers is exemplified. Through a modeling analysis by Matlab, setting and adjusting the unbalance penalty parameter, the relational graph about the change between total distance and unbalance measurement and several alternatives scored by an evaluation system are concluded. At last, an evaluation system is designed to choose the best answer, of which total distance is 40.819 5 and unbalance measurement is only one. The conclusion makes it clear that the model can help decision-making for managers. 

Key words: vehicle routing problem(VRP), load balancing, genetic algorithm, unbalance measurement 