摘要：

针对带硬时间窗车辆路径问题的多重模糊性，基于模糊可信性理论建立多目标模糊期望值模型，提出求解该问题的自适应混合多目标粒子群优化算法。该算法根据相位空间的思想给出一种实数编码方式，设计双存档机制，分别存储演化过程中产生的非支配解和有益不可行解，并引入自适应局部搜索、变异和粒子全局向导选择策略。仿真实验结果表明，与多目标进化算法相比，该算法可以获得更优的Pareto解集。

关键词: 车辆路径问题, 模糊可信性, 粒子群算法, 多目标优化, 约束, 时间窗, Pareto最优解

Abstract:

Aiming at the vehicle routing problem with hard time windows and multiple fuzzy characteristics, a multi-objective fuzzy expected model is designed based on fuzzy credibility theory, and an adaptive hybrid Multi-objective Particle Swarm Optimization(MOPSO) is proposed to solve the fuzzy vehicle routing model. The algorithm puts forward a particle encoding method according as phase-space, and designs a double archiving mechanism which stores the non-dominated solutions and excellent infeasible solutions separately. It also introduces adaptive strategies on local search, mutation and selection for particle global guide. The compareative experiments with multi-objective evolutionary algorithm verify that the method is capable of getting more excellent Pareto sets.

Key words: vehicle routing problem, fuzzy credibility, particle swarm algorithm, multi-objective optimization, constraint, time windows, Pareto optimal solution

中图分类号: 

• N945

• TP29