面向双层次选址路径问题的多任务强化演化优化方法研究

Research on Multi-task Reinforcement Evolutionary Optimization Method for Two-echelon Location-routing Problem

为提升双层次物流配送中心的选址和运输路径的整体优化性能,将双层物流选址路径问题看成是2个层不同的路径优化任务,提出一种多任务强化演化学习的双层次物流选址路径优化方法。首先,采用强化学习分别估计两层选址-路径问题中的上下节点分配选址概率;然后,设计基于分配概率的多任务交叉策略,并采用多因子演化算法协同优化不同层的物流选址路径,优化双层物流系统的成本。实验结果表明,提出的算法在求解双层物流选址路径优化问题上具有一定的优越性。

To enhance the overall optimization performance of two-echelon logistics distribution centers in terms of site selection and transportation routing,a multi-task reinforcement evolutionary learning approach is proposed for optimizing the two-echelon logistics site selection and routing problems.Firstly,reinforcement learning is employed to estimate the node allocation probabilities for site selection in each level of the problem.Then,a multi-task crossover strategy based on allocation probabilities is designed,and a multi-objective evolutionary algorithm is utilized to collaboratively optimize the logistics site selection and routing at different layers to optimize the cost of the two-tier logistics system.Experimental results show that the proposed algorithm has certain superiority in solving the two-echelon logistics site selection and routing optimization problem.