Abstract: Based on joint picking up and delivery by company's vehicles and social vehicles and national policies of environmental protection,a two-stage integer programming model was developed considering the Electric Dynamic Demand based Vehicle Routing Problem in the context of Crowdsourcing (EDDVRP-CD) with the goal of minimizing the total cost of delivery.Factors of time-of-use electricity price,partial charging,soft time windows and dynamic demand were simultaneously considered.Aiming at quick response to dynamic demand,an Improved Adaptive Large-scale Neighborhood Search and Tabu Search algorithm (IALNS-TS) was proposed with new destroy operators and repair operators developed,and an acceleration strategy was designed.The IALNS-TS was compared with Adaptive Large-scale Neighborhood Search algorithm (ALNS) and Tabu Search algorithm (TS).Based on extensive experiments,it proved that IALNS-TS algorithm could quickly respond to dynamic demands and effectively reduce the total distribution cost.

Key words: crowdsourcing mode, time-of-use electricity price, electric vehicle routing problem, dynamic demand, improved adaptive large-scale neighborhood search and tabu search algorithm

摘要： 针对企业自有车辆和社会车辆共同取送货的场景,以及国家节能环保的政策背景,考虑分时电价、部分充电、软时间窗、以及动态需求等因素,以最小化配送总成本为目标,建立考虑众包场景的电动车动态需求车辆路径问题(EDDVRP-CD)的两阶段整数规划模型。考虑动态需求的时效性,设计了启发式算法——改进的禁忌自适应大规模邻域搜索算法（IALNS-TS）,增加了新的删除算子和修复算子,同时提出了加速策略。分别与两种算法——自适应大规模邻域搜索算法（ALNS）以及禁忌搜索算法（TS）进行对比,通过大量算例验证了IALNS-TS算法能够快速响应动态需求,并有效降低总配送费用。

关键词: 众包模式, 分时电价, 电动车车辆路径问题, 动态需求, 改进的禁忌自适应大规模邻域搜索算法