【作者】 李建婷；

【导师】 刘林忠； 李爽；

【作者基本信息】 兰州交通大学 ， 交通运输工程（专业学位）， 2022， 硕士

【摘要】 铁路空车调配的基本任务是把各路局或者各地区的剩余空车调配到空车不满足的路局或者地区,以实现各个路局都能按计划完成每天的装卸任务。而有效提高铁路空车的分配效率和运输效率是铁路运输中的关键所在,也是取得更大经济效益的重要手段之一,因此有必要对铁路空车调配问题进行深入研究。(1)本文在既有研究的基础上,结合相关的空车调配理论知识,主要从动态性和随机性两个方面展开研究,其中的动态性通过建立时空网络来描述,随机性主要由随机变量来表现。(2)文中引入了既有研究中确定性条件下的铁路空车调配问题模型,并对该模型进行了分析评价,主要从三个方面进行了评价:模型的复杂性、确定性环境下的局限性、目标函数单一性。(3)针对模型评价中的模型复杂性,本文通过改变决策变量来实现模型的简化,确定性环境下模型的决策变量为各路径每条路段的空车流量,为了简化模型,以路网中各个OD间路径的空车流量作为决策变量,相对应的约束条件和变量的数量都会有所简化,且模型的复杂性不会随着路网复杂性的增大而增大,简化的模型在解决空车调配问题时发挥至关重要的作用。(4)针对模型评价中确定性环境下建立模型带来的局限性,本文主要在简化模型的基础上加入随机变量,考虑随机变量为路径通过能力和站点中转能力两个随机变量,以路网中路径上实际通过车辆数小于该路径最大通过能力的概率最大和站点各阶段实际中转车辆数小于该站点中转能力的概率最大为目标,以收益标准要求、空车供给站各阶段供给能力等作为约束条件,建立铁路空车调配问题的随机相关机会规划模型,并设计以随机模拟、神经网络和遗传算法相结合的混合智能算法进行求解。同时设计案例对所建立的模型和设计的算法进行验证,并与传统的遗传算法进行了对比,以求证文中模型及算法的高效性和实用性。(5)针对模型评价中目标函数单一性,文中在考虑收益最大化的同时考虑了需求方对空车调配满意度最大化,针对此类“效益悖反”的目标函数,在以上简化模型的基础上进行改进,建立了双层规划模型,双层规划模型能够较好地实现收益与满意程度同时实现最大的目标。其中上层规划以空车调配收益最大为目标,下层规划以综合满意度最大为目标,并且在下层规划中引入“综合选择度”作为下层规划服务质量满意度的定量化表述,文中主要考虑影响“综合选择度”的因素为车辆所处的阶段、车辆所在的路径以及车辆的种类,并在此基础上设计遗传算法对模型进行求解。最后引入实际案例分析模型及算法的有效性和实用性,并且给出了不同条件下获得最优解的具体数据,从而验证研究方法高效性,同时说明了该章节研究思路的正确性。更多还原

【Abstract】 The basic task of railway empty car allocation is to allocate the remaining empty cars in various railway bureaus or regions to railway bureaus or regions that are not satisfied with empty cars,so that each railway bureau can complete the daily loading and unloading tasks as planned.Effectively improving the distribution efficiency and transportation efficiency of railway empty cars is the key point in railway transportation,and it is also one of the important means to obtain greater economic benefits.Therefore,it is necessary to conduct in-depth research on the allocation of railway empty cars.(1)On the basis of the existing research and combined with the relevant theoretical knowledge of empty vehicle allocation,this paper mainly conducts research from two aspects:dynamics and randomness.The dynamicity is described by establishing a spatiotemporal network,and the randomness is mainly represented by random variables..(2)The paper introduces the problem model of railway empty car allocation under certain conditions in the existing research,and analyzes and evaluates the model,mainly from three aspects: the complexity of the model,the limitation in the deterministic environment and the unity of the objective function.(3)In view of the complexity of the model in the model evaluation,this paper simplifies the model by changing the decision variables.The decision variable of the model in the deterministic environment is the empty traffic flow of each road section of each route.The empty traffic flow of the route is used as a decision variable,and the complexity of the model does not increase with the complexity of the road network,and the simplified model plays a crucial role in solving the problem of empty vehicle allocation.(4)Aiming at the limitations brought by the establishment of the model in the deterministic environment in model evaluation,this paper mainly adds random variables on the basis of the simplified model,and considers the random variables as two random variables of route passing capacity and station transfer capacity.The maximum probability that the actual number of passing vehicles is less than the maximum passing capacity of the route and the maximum probability that the actual number of transit vehicles at each stage of the site is less than the transit capacity of the site are the goals.A stochastic correlation chance programming model for railway empty-car allocation problem is established,and a hybrid intelligent algorithm combining stochastic simulation,neural network and genetic algorithm is designed to solve it.At the same time,a case is designed to verify the established model and the designed algorithm,and compared with the traditional genetic algorithm to verify the efficiency and practicability of the model and algorithm in this paper.(5)Aiming at the singularity of the objective function in the model evaluation,the paper considers the maximization of the demand side’s satisfaction with the deployment of empty vehicles while considering the maximization of profits.Improvement,a two-level planning model is established,and the two-level planning model can better achieve the benefit and satisfaction degree and achieve the maximum goal at the same time.Among them,the upper-level planning aims to maximize the profit of empty vehicle allocation,and the lower-level planning aims to maximize the comprehensive satisfaction.In the lower-level planning,the "comprehensive selection degree" is introduced as a quantitative expression of the service quality satisfaction of the lower-level planning.The paper mainly considers the impact "" The factors of "comprehensive selectivity" are the stage of the vehicle,the path of the vehicle and the type of the vehicle.On this basis,a genetic algorithm is designed to solve the model.Finally,the validity and practicability of the actual case analysis model and algorithm are introduced,and the specific data of obtaining the optimal solution under different conditions are given,so as to verify the efficiency of the research method and illustrate the correctness of the research idea in this chapter.更多还原