混合算法求解时变路网中车辆调度规划问题

发布时间：2018-05-08 06:20

本文选题：物流规划 + 车辆调度　；参考：《上海交通大学》2014年硕士论文

【摘要】：随着运输物流业规模的扩大及成本控制需求的增加，对于运输车辆调度安排的优化也变得日益重要。良好的调度规划不仅能够极大程度地降低运输成本，同时也能有效地缩小到货所需时间，从而增加客户信赖度及黏度。然而伴随着城市化进程的日益推进，城市路网变得越发复杂，不同道路类型上的运输速度也大相径庭，这一速度不恒定因素使路网具有了时变性，也因此为车辆调度安排增加了复杂度与多样性。同时在实际物流的运输配送过程中还需要考虑运输车辆本身的载货量限制以及配送中心与配送点各自的营业时间窗，这些限制与时变性为车辆调度规划带来了难度与挑战。本文提出一个考虑车辆装载量容量约束、不同道路类型上速度时变的函数描述与时间求解、每个配送点包含有各自的营业时间窗约束条件，并分级考虑应用车辆配置规划与营运时间经济成本的符合现实的问题模型。并继而提出一个更适用的算法，，从而达到在更短时间里求得近似最优解的目的。本文提出的算法是将两个启发式算法相融合的混合算法，在多蚁群算法与遗传算法的基础上分别进行算法的改进再将两者融合，利用两者各自的优势来弥补对方在算法运行时的缺陷，从而实现寻解效率更高的目标。最后本文对实际交通路网情况进行了仿真实验，结果表明混合算法有效地弥补了多蚁群算法过早收敛于局部最优解而使算法停滞的缺陷，也改善了遗传算法收敛效率较慢，易进行冗余迭代的不足，从而在运行效率及寻优能力上比两种原算法都有所提高。
[Abstract]:With the expansion of the scale of transportation logistics industry and the increase of cost control demand, the optimization of vehicle scheduling becomes more and more important. Good scheduling planning can not only greatly reduce the transportation cost, but also effectively reduce the time required to arrive goods, thus increasing the customer trust and viscosity. However, with the development of urbanization, the urban road network becomes more and more complex, and the speed of transportation varies greatly with different road types. This unsteady speed factor makes the road network time-varying. Therefore, the complexity and diversity of vehicle scheduling are increased. At the same time, in the process of transportation and distribution of actual logistics, it is also necessary to consider the load limit of the transport vehicle itself and the business hours window of the distribution center and the distribution point. These limitations and time-varying bring difficulties and challenges to the vehicle scheduling planning. In this paper, a function description and time solution of time-varying speed on different road types considering the capacity constraint of vehicle loading capacity is proposed. Each distribution point has its own operating time window constraint condition. A realistic problem model considering the economic cost of vehicle configuration planning and operation time is considered. Then a more suitable algorithm is proposed to obtain the approximate optimal solution in a shorter time. The algorithm proposed in this paper is a hybrid algorithm which combines the two heuristic algorithms. Based on the multi-ant colony algorithm and genetic algorithm, the algorithm is improved, and then the two algorithms are fused. Using their respective advantages to make up for the other side in the algorithm running defects, so as to achieve a higher resolution efficiency goal. Finally, the simulation results of the actual traffic network show that the hybrid algorithm can effectively compensate for the premature convergence of the multi-ant colony algorithm to the local optimal solution and the stagnation of the algorithm, and improve the slow convergence efficiency of the genetic algorithm. It is easy to carry out redundant iteration, so the efficiency and optimization ability of the two algorithms are improved.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U492.22;TP18

【参考文献】