Multi-Scenario Capacitated Vehicle Routing Optimization Algorithm Based on Parallel Evolution Framework

多场景优化问题涉及需要同时优化的多个场景，每个场景都具有特定条件和多个待优化目标。其目标在于寻找一组公共折衷最优解(PCOS)：这些解在所有场景中均可行，且能达成折中的优化效果。带容量约束的多场景车辆路径问题(MSCVRP)考虑了现实配送中的不确定扰动，目前针对该问题研究较少。本文提出一种基于并行进化框架的多场景带容量车辆路径优化算法(PEF-MSOA)。通过利用各场景的knee点和约束极值线构建不同场景间解的转移策略，实现不同场景间的有效信息传递。此外，设计自适应算子选择策略以提升PCOS解集的质量。对比实验中验证了所提算法的有效性与优越性。Multi-scenario optimization problems involve multiple scenarios that need to be optimized simultaneously, each with specific conditions and multiple objectives to optimize. The goal is to find a set of public compromised optimal solutions (PCOS): feasible in all scenarios and achieving a balanced optimization effect. The multi-scenario capacitated vehicle routing problem (MSCVRP) considers uncertainty disturbances in real-world delivery, with few effective measures for this problem in existing research. In this paper, a multi-scenario capacitated vehicle routing optimization algorithm based on the parallel evolution framework is proposed. The knee solutions and extremal line of each scenario are used to construct a solution transfer strategy, thus achieving information transfer between different scenarios. Besides, an adaptive operator selection strategy is designed to improve the quality of PCOS. The effectiveness and superiority of the proposed algorithm are verified through comparative experiments.