【作者】 张凡；

【导师】 刘志硕；

【作者基本信息】 北京交通大学 ， 物流工程（专业学位）， 2020， 硕士

【摘要】 随着我国城市化进程的不断加快,城市生活垃圾的产生量也与日俱增,对城市生活垃圾进行分类处理显得尤为重要。分类收运是生活垃圾分类处理过程中不可或缺的重要环节。然而,目前普遍采用燃油车进行城市生活垃圾分类收运无疑增加了化石能源消耗和环境负担。随着电动汽车的发展,采用电动汽车代替燃油车是城市生活垃圾分类收运的发展方向。此外,考虑到社区等生活垃圾点同时拥有多种垃圾,采用带隔舱的车辆能够大幅降低垃圾点的访问次数,从而有效降低行驶里程。因此,本文尝试将带隔舱的电动汽车应用于城市生活垃圾分类收运领域,提出带隔舱电动汽车的生活垃圾分类收运车辆路径问题,建立数学规划模型并设计一种混合蚁群算法求解,以为城市生活垃圾分类收运的车辆调度提供智能辅助决策支持。本文研究工作主要如下:本文首先对城市生活垃圾分类收运业务进行系统分析。根据有无中转站,将城市生活垃圾分类收运业务划分为直接收运模式和间接收运模式。在目前普遍采用的间接收运模式中,将生活垃圾从社区等垃圾点分类收运到中转站是整个收运过程的开始环节。针对该环节的车辆调度业务需求,本文提出了城市生活垃圾分类收运电动车辆路径问题(Multi-Compartment Multi-Trip Heterogeneous-Fleet Electric Vehicle Routing Problem with Hard Time Window,E-MMHVRPHTW),并在该问题中考虑了硬时间窗、多种车型、车辆可以连续运行多趟、车辆带隔舱。然后在考虑车辆容量约束、车辆电能约束和时间窗约束的基础上构建了以车辆固定成本以及车辆能耗成本之和最小为目标函数的数学规划模型。设计了分别融合4种局部优化算子的混合蚁群算法求解E-MMHVRPHTW问题,该算法能够让蚂蚁模拟车辆从中转站出发,根据转移规则不断访问各个垃圾点或社会充电站,完成路径的构造,形成问题的解。并基于正反馈机制,采用精英蚂蚁策略,根据解的质量进行信息素的更新,不断迭代,最终得到优质的调度方案。最后基于Solomon算例构建了适用于E-MMHVRPHTW问题的算例,形成了小规模和大规模算例集,并通过大量实验来验证混合蚁群算法的性能,实验结果表明,混合蚁群算法比单纯蚁群算法的性能更好,尤其体现在求解大规模算例上。并将混合蚁群算法应用到垃圾分类收运实例中,验证了该算法具有一定的实用性。更多还原

【Abstract】 With the acceleration of urbanization in China,the quantity of municipal solid waste is increasing gradually,and it is very important to classify and dispose the municipal solid waste.Classified pickup and transportation is an indispensable part in the process of municipal solid waste classification and disposal.However,the current widespread use of fuel vehicles for the classified pickup and transportation of municipal solid waste has undoubtedly increased fossil energy consumption and environmental burden.With the development of electric vehicles,the use of electric vehicles instead of fuel vehicles is the development direction of the classified pickup and transportation of municipal solid waste.In addition,considering that household garbage sites such as communities have multiple kinds of garbage at the same time,the use of vehicles with compartments will greatly reduce the number of garbage points visits,thereby effectively reducing the mileage.Therefore,this paper attempts to apply the electric vehicles with compartments to the field of municipal solid waste classification,pickup and transportation,proposes the routing problem of electric vehicles with compartments for the classified pickup and transportation of household garbage,establishes a mathematical programming model and designs a hybrid ant colony algorithm.It can provide intelligent auxiliary decision support for vehicle scheduling of municipal solid waste sorting and transportation.The main research work of this paper is as follows:This paper firstly conducts a systematic analysis on the classified pickup and transportation business of municipal solid waste.According to whether there is a transfer station,the classified pickup and transportation business of municipal solid waste is divided into a direct pickup and transportation mode and an indirect pickup and transportation mode.In the currently accepted indirect pickup and transportation mode,it is the beginning of the whole pickup and transportation process to collect and transport household garbage separately from the community and other garbage points to the transfer station.Aiming at the vehicle scheduling business requirements in this link,this paper proposes a routing problem for the classified pickup and transportation of municipal solid waste based on electric vehicles.In this problem,it is considered that the hard time window,heterogeneous fleet,vehicles can run continuously for multiple times and vehicles with compartments.Then based on the consideration of vehicle capacity constraints,vehicle electrical energy constraints and time window constraints,a mathematical programming model with the minimum sum of vehicle fixed cost and vehicle energy consumption cost as the objective function is constructed.A hybrid ant colony algorithm combining four local optimization operators respectively is designed to solve the E-MMHVRPHTW problem.The algorithm enables ants to simulate vehicles starting from the transfer station,and continuously visit each garbage point or social charging stations according to the transfer rule to complete the path construction and form a solution to the problem.And based on the positive feedback mechanism,the elite ant strategy is used to update the pheromone according to the quality of the solution,and iterate continuously to finally obtain a highquality scheduling plan.Finally,based on the Solomon benchmark example,two sets of small-scale and large-scale examples for the E-MMHVRPHTW problem are formed.The performance of the hybrid ant colony algorithm is verified through a large number of experiments.The experimental results show that hybrid ant colony algorithm has better performance than simple ant colony algorithm,especially in solving large-scale examples.The hybrid ant colony algorithm is applied to the example of garbage pickup and transportation,which verifies the practicality of the algorithm.更多还原