Abstract: To meet the increasingly diversified and customized demands and to reduce the costs of inbound logistics in the automotive industry,the joint optimization of auto parts distribution inventory routing was studied.The characteristics of auto parts including time-varying demands deriving from the mixed-model assembly line and the fact that the auto parts should be delivered in original packages provided by the suppliers were considered to construct an integer linear programming model by making decisions on multi-stage inventory level and vehicle take delivery time,and the feasibility conditions were analyzed.A bi-level genetic algorithm was designed to solve the problem.In each iteration of the proposed algorithm,an upper level sub-algorithm determined the reorder point and the maximum inventory level for each type of the auto parts,and a lower level sub-algorithm solved the routing problem for the trucks,where some hyperparameters were predetermined by experiments to ensure the quality of the solutions and to reduce the computational time.Computational results based on randomly generated instances showed that the proposed algorithm was superior in solving the model directly by commercial solver CPLEX,in terms of both solution quality and computation time.The proposed model and algorithm would provide scientific and effective supports on inbound logistics operations for automotive manufacturers and third-party logistics companies.

Key words: auto parts, inbound logistics, inventory routing, integer programming, bi-level genetic algorithm

摘要： 为满足现阶段日益多样化、个性化的汽车市场需求,进一步降低汽车零部件入厂物流成本,针对零部件配送库存路径联合优化问题进行研究。以多阶段库存零部件水平和物流车辆取货时序为决策,考虑生产线的时变需求和零部件须按供应商提供的原包装进行整包配送的实际特征,建立了汽车零部件入厂物流库存路径问题的线性整数规划模型,并分析了可行性条件。根据问题特征,设计了双层遗传算法进行求解:上层算法求解再订货点和最大库存量,下层算法求解每辆卡车的循环取货任务和路径,其中通过实验提前确定了可变超参数以保证解的质量同时减少计算时间,上下层算法重复迭代直至求得近似最优解。基于随机算例的数值实验结果表明,随着算例规模增长,双层遗传算法在求解质量和时间上均优于商业求解器CPLEX直接求解数学模型,该模型和算法可以为汽车制造商及第三方物流提供科学高效的零部件入厂物流的决策支持。

关键词: 汽车零部件, 入厂物流, 库存路径, 整数规划, 双层遗传算法