Abstract: In order to reduce the cost of land-use and inventory, many manufacturers replace the traditional warehouse with limited shipping area nearby the production line. The finished products transported to limited capacity shipping area are allowed to stay for a short time before being transported to central products warehouse by a fleet of vehicles. This so called line-pickup mode leads challenge for vehicle scheduling since the overflow of shipping area is prohibited. Inspired by inventory discretization idea, the continuous arrival of inventory at a plant can be split into a sequence of discrete tasks for pickup, so that the problem can be converted into multiple-trip vehicle routing problem with time window. The modified PSO algorithm is used, while clustering and classification are proposed to improve algorithm’s efficiency. Finally, a practical case is studied to demonstrate and validate the developed approach.

Key words: line-pickup, inventory discretization, multiple-trip vehicle routing problem, Particle Swarm Optimization（PSO）

摘要： 为了节约仓储用地和库存成本，许多制造企业取消了传统的仓库而采用了在工厂生产线边设置面积有限的发货区。成品运送至有限发货区后，只做短暂停留就被第三方物流运往中央成品仓库。这种线边接货模式给车辆运输调度带来了困难，因为需要保证发货区的货物临时存储不能溢出。借助于发货区库存离散化的思想，可将进入发货区的货物分拆成若干个运输任务，进而将上述线边接货模式下的车辆运输调度优化问题转化带时间窗的多趟车辆路径优化问题，进而利用改进的粒子群算法求解，采用聚类和划分任务等级来提高算法的效率。通过工厂实际案例对算法的可用性和有效性进行验证。

关键词: 线边接货, 库存离散化, 多趟车辆路径问题, 粒子群算法