考虑多车型软时间窗的成品油二次配送库存-路径问题

Inventory Routing Problem of Refined Oil Secondary Distribution Considering Multi-vehicle and Soft Time Windows

针对成品油二次配送库存-路径问题(inventory routing problem,IRP),考虑了随机需求、多车型多隔舱等因素,以车辆配送成本、违反时间窗等惩罚成本之和最小化为目标,建立了混合整数规划模型。设计了求解模型的两阶段启发式算法,第一阶段根据各加油站的期初库存量、销量的概率分布等确定补货量和时间窗;第二阶段利用遗传算法和C-W节约算法确定配送车辆和配送路径。通过模拟计算并与单阶段算法和Gurobi直接求解模型的结果进行对比,验证了两阶段算法的有效性。结果显示利用本文方法确定成品油二次配送计划,在避免加油站缺货的同时,有效降低了车辆的配送成本。

The inventory-routing problem of refined oil secondary distribution was studied,considering the factors consist of stochastic demand,multi-vehicle and multi-compartment.In order to minimize the total cost such as distribution cost and the penalty cost for violation of time windows,a mixed integer programming model was established.A two-stage heuristic algorithm was designed.In the first stage,the replenishment quantity and time windows were determined according to the initial inventory,probability distribution of sales and so on.In the second stage,the distribution routes and each vehicle were obtained by the genetic algorithm and C-W algorithm.Numerical experiments and results comparison of single-stage algorithm and Gurobi solver verify the effectiveness of the algorithm proposed.The results indicate that our method can not only avoid the stockout of petrol stations,but also effectively reduce the distribution routing cost.