基于多目标优化算法的公交车辆调度研究

发布时间：2018-04-16 06:44

本文选题：车辆调度问题 + 公共交通　；参考：《北京邮电大学》2014年硕士论文

【摘要】：伴随着我国社会的不断发展,越来越多的人民群众选择公共交通作为首选的出行工具。同时大力发展公共交通,可以有效缓解城市道路堵塞的状况,并对解决空气污染问题有很大的帮助。在公共交通运行体系当中,公交车辆调度问题是一个非常复杂并且困难的多目标优化问题。好的公交调度方案不仅能够减少公交公司的运营成本,而且能够大大提高公共交通的效率,减少乘客候车时间,鼓励更多人选择公共交通出行。公交车辆调度问题是指调配车辆来覆盖一系列包含在公交发车时刻表中的行程,同时最小化多个目标函数,比如使用的车数和司机数,这多个目标之间通常是互相冲突的。当前已知的方法是将多个目标通过线性组合转化成单个目标,然后采用单目标优化方法来求解。然而,这种方法只能够产生一个解,并且很难为每一个目标函数分配一个合适的权重值,从而得到一个平衡各个目标的最优解。本文调研了我国公交车辆调度的实际情况,并在借鉴了车辆调度问题相关研究的基础上,提出了一种多目标优化方法来解决公交车辆调度问题。该方法能够产生多个Pareto最优解,每个Pareto最优解代表一种符合要求的车辆排班方案。该方法的主要步骤如下：(1)产生一系列初始行程块集合,作为候选集合；(2)提出从候选集合中选择子集构成车辆排班方案的多目标优化模型,该模型包含车辆数和司机数两个目标；(3)采用一种改进的多目标遗传算法来对模型进行求解,得到多个Pareto最优解。提出的编码方案可以有效地减少编码长度和解码的复杂度；(4)使用发车时间调整过程(DTAP)来提高解的质量。为了验证方法的有效性,本文使用该方法来解决南京公交公司实际公交线路的车辆调度问题。实验表明该方法可以在十几分钟内生成多个符合要求的排班方案,并且优于实际使用的人工排班方案。
[Abstract]:With the development of our society, more and more people choose public transportation as the first choice of travel tools.At the same time, the development of public transport can effectively alleviate the urban road congestion and solve the air pollution problem.In the public transportation system, the bus scheduling problem is a very complex and difficult multi-objective optimization problem.A good bus scheduling scheme can not only reduce the operating costs of public transport companies, but also greatly improve the efficiency of public transport, reduce the waiting time of passengers, encourage more people to choose public transport travel.Bus scheduling problem refers to the deployment of vehicles to cover a series of trips included in the bus timetable, while minimizing multiple objective functions, such as the number of vehicles used and the number of drivers, which are usually in conflict with each other.The current known method is to transform multiple targets into a single target by linear combination, and then to solve the problem by using a single objective optimization method.However, this method can only produce one solution, and it is difficult to assign an appropriate weight value for each objective function, thus an optimal solution to balance each objective is obtained.This paper investigates the actual situation of public transport vehicle scheduling in our country, and proposes a multi-objective optimization method to solve the bus vehicle scheduling problem on the basis of referring to the related research of the vehicle scheduling problem.The method can generate multiple Pareto optimal solutions, and each Pareto optimal solution represents a vehicle scheduling scheme that meets the requirements.The main steps of this method are as follows: (1) A series of initial travel block sets are generated as candidate sets. (2) A multi-objective optimization model is proposed to select subsets from candidate sets to form a vehicle scheduling scheme.The model consists of two targets: the number of vehicles and the number of drivers. An improved multi-objective genetic algorithm is used to solve the model and multiple Pareto optimal solutions are obtained.The proposed coding scheme can effectively reduce the coding length and decode complexity.In order to verify the effectiveness of the method, this paper uses this method to solve the actual bus routing problem of Nanjing bus Company.The experimental results show that this method can generate several scheduling schemes in ten minutes, and it is superior to the actual manual scheduling scheme.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U492.22;TP18

【共引文献】