基于深度强化学习的带约束车辆路径分层优化研究

打开文本图片集

中图分类号：TP301 文献标志码：A DOI：10.12305/j.issn.1001-506X.2025.03.15

Hierarchical optimization research of constrained vehicle routing based on deep reinforcement learning

TANG Kaiqiang，FU Huiqiao，LIU Jiasheng，DENG Guizhou，CHEN Chunlin （Schoolof EngineeringManagement，NanjingUniuersity，Nanjing 2loo93，China）

Abstract：For the capacitated vehicle routing problem （CVRP），a method is proposed to decouple the capacityconstraints using a hierarchical structure，split the complex CVRP into constraint planning and path planning，and perform deep reinforcement learning（DRL）optimisation for solving the problem respectively. Firstly，the upper layer alocates the vehicle distribution tasks based on the atention model and sampling mechanism to plan the set of subpaths that satisfy the constraints. Secondly，the lower layer adopts the pretrained unconstrained atention model to plan the paths for the setof subpaths.Finally，the network parameters of the upper layer are optimized through the feedback training and iteration of the Reinforce algorithm. Experimental results show that the method generalizes to CVRPand heterogeneous CVRP tasks of diferent sizes， outperforms thestate-of-the-art DRL method.Moreover，compared with other heuristic methods，in batch computing tasks，the solution speed improved by more than 1O times，while maintaining competitive solutions.

Keywords：deep reinforcement learning （DRL）；vehicle routing problem（VRP）；attention model；hierarchical optimization

0 引言

二[1-3]，广泛存在于物流、工业和运输等多个领域。（剩余21215字）