双重信息引导的蚁群算法求解绿色多舱车辆路径问题

Dual-information guided ant colony optimization algorithm for green multi-compartment vehicle routing problem

中文关键词  多舱车辆路径问题  绿色  蚁群优化算法  双重信息引导  信息素浓度平衡机制

英文关键词  multi-compartment vehicle routing problem  green  ant colony optimization algorithm  dual-information guided  pheromone concentration balance mechanism

中文摘要

      针对当前实际运输中广泛存在的绿色多舱车辆路径问题(GMCVRP), 文章提出一种双重信息引导的蚁群优化算法(DIACO)进行求解. 首先, 在DIACO的全局搜索阶段, 重新构建传统蚁群优化算法(TACO)中的信息素浓度矩阵(PCM), 使其同时包含客户块信息和客户序列信息, 即建立具有双重信息的PCM(DIPCM), 从而更全面学习和累积优质解的信息; 采用3种启发式方法生成较高质量个体, 用于初始化DIPCM, 可快速引导算法朝向解空间中优质区域进行搜索. 其次, 在DIACO的局部搜索阶段, 设计结合自适应策略的多种变邻域操作, 用于对解空间的优质区域执行深入搜索. 再次, 提出信息素浓度平衡机制, 以防止搜索陷入停滞. 最后, 使用不同规模的算例进行仿真测试和算法对比, 结果验证了DIACO是求解GMCVRP的有效算法.

英文摘要

      For dealing with the green multi-compartment vehicle routing problem (GMCVRP) widely existing in actual transportation, a dual-information guided ant colony optimization algorithm (DIACO) is proposed to solve it. First, in the global search stage of DIACO, the pheromone concentration matrix (PCM) in the traditional ant colony optimization algorithm (TACO) is reconstructed. The reconstructed PCM contains both customer block information and customer sequence information. That is, the dual-information PCM (DIPCM) is established so as to more comprehensively learn and accumulate high-quality solution information. Three types of heuristic methods are adopted to generate higher quality individuals for initializing DIPCM, which can guide the algorithm to search for high-quality regions in the solution space quickly. Second, in the local search stage of DIACO, multiple variable neighborhood operations combined with adaptive strategy are designed to perform in-depth searches on high-quality regions of the solution space. Third, the pheromone concentration balance mechanism is proposed to prevent the search from stagnating. Last, simulation tests and algorithm comparisons are carried out with different scale examples. The results show that DIACO is an effective algorithm for solving the GMCVRP.