导　　师： 温惠英

学科专业： 082303

授予学位： 硕士

作　　者： ;

机构地区： 华南理工大学

摘　　要： 协同车辆路径问题（Collaborative Vehicle Routing Problem，CVRP）是多配送中心车辆路径问题/(Multi-depot VRP，MDVRP/)发展的高级阶段，研究各物流企业通过Internet等信息技术创造协同环境，共享各种信息和资源，如顾客、仓库和车队等，为所属不同公司的客户统一装卸货物以降低物流配送成本。CVRP是物流方面的一个新研究领域，目前国内外相关理论研究较少。 本课题以CVRP为研究对象，主要工作如下： （1）在满足最大行驶里程、车辆容量限制、时间窗等现实约束情形下，以车辆配送总费用最小为目标，研究一类确定性CVRP，设计相应的离散粒子群算法，通过综合仿真和试验分析验证模型的正确性和合理性。仿真表明：同普通物流配送情形相比，CVRP有效减少总配送里程和配送费用。 （2）考虑不确定因素干扰CVRP，研究一类模糊旅行和顾客服务时间的CVRP，构建了该问题的模糊规划模型，并将之进行清晰化处理使之转换为一类确定性数学模型，设计相应的离散粒子群算法等。在突发事件干扰下，该研究编制的配送方案可以适应不断变化的交通环境，从而更加贴近物流企业的实际应用。 （3）考虑车辆依原计划配送过程中的随机干扰因素导致配送方案失效，引入可靠性理论探讨CVRP，定义了车辆任务可靠度概念，在此基础上研究一类基于车辆任务可靠度的CVRP。根据问题特征，设计求解该CVRP问题的遗传算法，定义了解的编码方案、产生初始种群的启发式算法等，通过综合仿真和试验分析验证了模型的正确性和合理性。研究表明：随着车辆任务可靠度的增大，物流配送方案费用将可能越高，但该方案的可靠性较好，故它在物流实际配送中受不确定性的干扰较小。 本课题研究成果丰富了车辆路径问题的内容和形式，既填补了CVRP领域的� Collaborative Vehicle Routing Problem /(CVRP/) is the advanced stage of development ofthe multi-depot Vehicle Routing Problem /(MDVRP/). The various logistics companies shareinformation and resources such as customers, warehousing and motorcade through internet.They serve the customers belong to different companies to reduce the cost of total distribution.CVRP is a new research field of Logistics, and the related.theoretical study is rarely. The paper aims at studying CVRP, and the main research work are summarized asfollows: /(1/)Considering some constraints such as maximum mileages of different vehicles,depot’s capacity and time window, etc., a class of deterministic CVRP to optimize total cost ofdispatching was studied. The problem was solved by discrete particle swarm optimizationalgorithm, and the model’s correctness and reasonableness was proved throughComprehensive simulation and experimental analysis. The experimental result shows that theoptimal total distribution mileage and cost solved by the presented model can be significantlyreduced compared to general logistics. /(2/) A class of CVRP with fuzzy travel and customer service time was discussed. Thispaper built a fuzzy programming model on CVRP，considering uncertainty factors to affectCVRP, and then converted into a deterministic one. Its solution was also solved by discreteparticle swarm optimization algorithm. The distribution of the prepared program can adapt tochanging traffic environment, when the incident interference occurred. Therefore, it is closerto the practical application of logistics enterprises. /(3/) Considering the random factors in the distribution of the original plan, which resultedin a failure of distribution program, reliability theory is used in the CVRP in the paper. AConcept to the vehicle task reliability was defined. Based on feature on the problem, solutionalgorithm was designed. Defining scheme of the solution code and the heuristic algorithm toproduce the initial population, the model’s correctness

关 键 词： 物流 协同运输 车辆路径问题 粒子群算法 遗传算法

领　　域： [交通运输工程]