Abstract:

Under the JIT procurement mode, the influence of transportation cycle and purchase quantity on path optimization is studied, the scheduling model is established, and a grid-based adaptive Adaptive Artificial Bee Colony Algorithm （GAMOABC） is designed to minimize the total transportation distance and the number of vehicles used. In the algorithm, the pareto optimal solution saved in the grid is used to update and maintain the optimal solution in the grid, so as to ensure the diversity of solution set and update the leading bee position in the grid through location sharing information, so as to improve the accuracy of solution set. The priority value corresponding to the vehicle and raw materials is represented by the encoding method of two-dimensional matrix. In the decoding process, in order to meet the production constraints, the scheduling set is determined according to the current raw material consumption completion time, and the heuristic information is designed. The test examples and experiments show that compared with NSGA-II and MOEAS, the Pareto solution set obtained by GAMOABC algorithm is more diverse and accurate.

Key words: JIT purchasing, artificial bee colony algorithm based on adaptive grid, multi-objective vehicle scheduling, transportation economy

摘要：

在JIT采购模式下，以最小化采购运输总距离和车辆使用数目为双目标，重点研究了运输周期和采购量对路径优化的影响，建立了调度模型，设计了一种基于自适应网格的多目标人工蜂群算法（Grid-based Adaptive Multi-Objective Artificial Bee Colony Algorithm，GAMOABC）。算法中，利用网格保存找到的帕累托最优解，对网格内的最优解进行更新和维护，保证解集的多样性并通过位置共享信息，更新网格内引领蜂的位置，从而提高解集的精确性。利用二维矩阵的编码方式表示车辆与原料对应的优先权值。在解码过程中，为满足生产约束，根据当前原料的消耗完成时间确定调度集合，设计了启发式信息。通过测例及实验表明：相较于NSGA-II、MOEAS算法，GAMOABC算法求得的Pareto解集多样性和精确性更好。

关键词: JIT采购, 自适应网格的人工蜂群算法, 多目标车辆调度, 运输经济