【作者】 严明；

【导师】 刘鸿雁；

【作者基本信息】 辽宁科技大学 ， 计算机应用技术， 2010， 硕士

【摘要】 随着我国铁路现代化建设的高速发展,车站编组调度效率瓶颈已凸显。作为一类典型组合优化复杂问题,货运列车编组调度优化成为具有理论水平和实际应用意义的重要研究领域。本文介绍了现阶段铁路信息化的成果与不足、国内外对于货运列车编组调度模型与算法研究现状。阐述了Agent、MAS的相关概念、性质及应用等基础知识。分析了几种经典优化理论方法。研究了博弈论的发展、类型划分、表示形式及算法实现平台等。以GMAIM为原型,进一步研究应用于编组调度问题的多Agent博弈协作,建立了货车编组调度博弈概念模型,并设计了相应博弈算法。根据货运列车编组调度过程的性质和特征,分析讨论了货车编组调度的工作流程、效率评价指标、优化目标及约束条件,建立了基于博弈论的货运列车编组调度优化可解模型。本文比较了多种常用的系统作业调度算法。提出了针对编组调度模型的编组调度贪心算法、基于网络优化配流的表上作业法、编组调度遗传算法等3类求解算法。作者研究了基于FCFS的货车编组调度算法实现。以某编组站为计算实例,针对3类编组调度优化算法进行了算法效率比较,并分析了中时上限、满轴率等因素对编组调度效率的影响。最后,作者分析了影响编组站解体能力的相关因素。在参考现存研究成果的基础上,建立了编组站通过能力计算模型,针对性的研究了编组站设备数量与负荷优化策略。通过编组站到解过程仿真数据的构建,在设置一系列的初始仿真条件下对编组站到解作业过程进行连续多次仿真。更多还原

【Abstract】 With the rapid development of the modernization of China’s railway, the efficiency of marshalling stations is an obvious bottleneck, as a typical combinatorial-optimizing complex problem, the optimization of marshalling-scheduling of freight trains attracts the attentions of many countries and scholars to be an academic hot-spot.The paper describes the advantages and shortcomings of railway informational at this stage and the researches of the marshalling-scheduling model and algorithm for the freight train at home and abroad. The author describes the basic knowledge, as the related concepts, nature and application of Agent, MAS. Several classical optimal theory methods are analysis of. The development, type, classification and representation of game theory and the platform of algorithm implementation have been studied.Taking the GMAIM as a prototype, the author makes further research in marshalling-scheduling problem of multi-Agent Game collaboration, and establishes a marshalling-scheduling game-concept model of freight trains, designs the corresponding game algorithm.According to the nature and characteristics of freight train marshalling-scheduling process, the author analyzes and discusses the marshalling-scheduling workflow, efficiency evaluation index, optimization objectives and constraints of the freight trains, establishes a solvable marshalling-scheduling optimal model of freight trains based on game theory.The author introduces a variety of common system job scheduling algorithms with comparing. 3 kinds of algorithms for salvation of the marshalling-scheduling model are proposed: the marshalling-scheduling greedy algorithm, the operating table method based on the network optimization, and the genetic algorithm.The author studies the achieve of marshalling-scheduling algorithms of freight trains based on FCFS. Taking a marshalling station as calculation example, for the three categories of marshalling-scheduling optimal algorithms the efficiency of the algorithms is compared, analyzing in the impact of the time-limit, full-axis rate and other factors on the marshalling-scheduling efficiency.Finally, the author analyses the relevant impact factors of the disintegration of the marshalling yard capacities. In reference to the existing research results, the marshalling yard capacity calculation model is established, targeted to the optimization strategy of the equipment quantity and load whithin the marshalling yard. By constructing the simulation data for the arriving-disintergration marshalling process, setting up a series of initial simulation, the author made simulation of the marshalling process for times.更多还原