带时间窗装卸一体化车辆路径问题的混合离散粒子群优化算法  

A Hybrid Discrete Particle Swarm Optimization Algorithm for Vehicle Routing Problem with Time Windows and Simultaneous Pickup and Delivery

作　　者：周蓉[1] 沈维蕾[1] 刘明周[1] 赵韩[1] 

机构地区：[1]合肥工业大学,合肥230009

出　　处：《中国机械工程》2016年第4期494-502,共9页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(71071046)

摘　　要：为了同时实现总配送成本最低、车辆数最少和车辆行驶距离最短等目标,考虑车辆指派成本及运输路径成本的相对重要性,建立了带时间窗装卸一体化车辆路径问题的混合整数规划模型。针对该问题搜索空间的离散性和求解算法的局部收敛性,提出了一种混合离散粒子群求解算法。算法基于客户排列的直观无分段大路径解表示法,采用改进深度优先搜索分割法对问题解进行解码与评价;嵌入一种变邻域下降搜索程序并在个体粒子每次迭代时以一定概率选择执行,利用混合粒子群算法在多邻域深度搜索和在全局空间广度搜索进行寻优,同时应用模拟退火思想和比例选择性变异最差个体来改善个体搜索停滞现象。采用两个不同目标算例进行寻优测试,验证了所提算法的可行性和有效性。Considering the relative importance of dispatching cost of vehicles and travelling cost of routes, a mixed integer mathematical formulation of the vehicle routing problem with time windows and simultaneous pickup and delivery （VRPTWSPD） was established to make the total delivery cost, the number of vehicles and the travel cost minimize simultaneously. A hybrid discrete particle swarm optimization algorithm was proposed for solving VRPTWSPD. Solutions of algorithm were represen- ted by an intuitive fragment-free giant tour of a permutation of all customers, and were decoded and e- valuated by the revised depth first search split procedure. A variable neighborhood descent search pro- cedure was carried out under a certain probability for individuals at each iteration, so that individuals could update with depth search in multi-neighborhood and with bread search in global exploration. The simulated annealing idea and mutating a selective portion of worst individuals were applied to im- prove the stagnation of the search. The feasibility and effectiveness of the proposed algorithm were verified by two instances with different goals.

关 键 词：带时间窗车辆路径问题 装卸一体化 离散粒子群优化算法 变邻域下降搜索 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]