• •    

行驶时间随机的分批配送车辆路径问题模型与算法研究

石建力1,张锦2   

  1. 1. 西南交通大学
    2. 西南交通大学 交通运输与物流学院, 成都 610031
  • 收稿日期:2017-07-31 修回日期:2017-09-28 发布日期:2017-09-28
  • 通讯作者: 石建力

Model and Algorithm for the Split Delivery Vehicle Routing Problem with Stochastic travel times

ZHANG Jin2   

  • Received:2017-07-31 Revised:2017-09-28 Online:2017-09-28

摘要: 摘 要: 本文研究行驶时间随机的分批配送车辆路径问题,在软时间窗下考虑等待时间,建立带修正的随机规划模型。设计改进的粒子群算法进行求解,对编码和解码过程、位移更新、位置更新和局部搜索等进行改进,将自适应选择和路径重连算法融入粒子群算法,以适应允许分批配送的特点。通过在调整的Solomon算例测试,考虑等待时间将造成总费用平均增加约3%,且更倾向于分批配送。分批配送能有效降低总费用和使用车辆数,分别平均降低2%和少0.6辆;在部分算例,特别是R2类算例中,分批配送能有效降低等待时间,平均降低0.78%。

关键词: 粒子群算法, 分批配送, 随机行驶时间, 车辆路径问题, 软时间窗

Abstract: Abstract: A stochastic programming with recourse was formulated to solve the split delivery vehicle routing problem with stochastic travel time, which considers the waiting time under soft time windows. An improved particle swarm optimization with adaptive selection rule and path relinking is developed for this problem, in which the coding and decoding process, the updating of the velicoty, the updating of the position and the local search process were designed for the split delivery factor. Computation tests on modified Solomon’s instances were carried out. The results showed that, the all-in cost increases about 3% averagely, and the splitted customers becomes more under the consideration of waiting time. The all-in cost decrease about 2% averagely and the number of vehicle decrease about 0.6 by allowing the split delivery. What’s more, the waiting time decreases 0.78% averagely because of the split delivery, especially in the instances in R2.

Key words: particle swarm optimization, split delivery, stochastic travel time, VRP, time windows