基于深度学习的离散型制造业库存管理与物流配送设计与优化

Design and Optimization of Inventory Management and Logistics Distribution in Discrete Manufacturing Based on Deep Learning

随着制造业的发展,离散型制造业库存管理和物流配送已成为制造业企业最为重要的管理领域。由于其高度复杂性,该领域面临着多种挑战,增加了企业的运营成本和管理难度。基于此,研究针对物流企业车辆调度活动中需要找出各个配货节点之间的最短路径,用以指导物流车辆调度的问题,研究提出一种将遗传算法(Genetic algorithms,GA)与BP神经网络算法(Back Propagation Neural Networks,BPNN)相结合的新方法。结果显示:改进后的GA算法在迭代次数较少的情况下,就可使平均种群适应度更靠近最大种群适应度;改进的GA-BPNN算法得到各项成本总计7635.14元,明显优于传统BPNN算法。该研究为制造业企业提供了一种新的车辆调度策略,可有效减少运输成本,提高物流服务水平。

With the development of manufacturing industry,inventory management and logistics distribution in discrete manufacturing industry has become the most important management area for manufacturing enterprises.Due to its high complexity,this area faces multiple challenges that increase the operational costs and management diffi culties of enterprises.Based on this,the study proposes a new method that combines genetic algorithms(GA)and BP neural networks(BPNN)to fi nd the shortest paths between distribution nodes to guide the dispatching of logistics vehicles in logistics enterprises,and proposes a new method combining Genetic algorithms(GA)and BP Neural Networks(BPNN).The results show that the improved GA algorithm can make the average population fi tness closer to the maximum population fi tness with fewer iterations,refl ecting that the improved algorithm eff ectively can improve the population fi tness capability.The improved GA-BPNN algorithm obtains a total cost of$7635.14 for each item,which is significantly better than the traditional BPNN algorithm.This study can provide a new vehicle scheduling strategy for manufacturing enterprises,which will eff ectively reduce transportation costs and improve logistics services.