摘要：本论文针对轨道交通中列车运行过程模型和节能操纵优化进行了研究。运行过程优化就是从众多的操纵方法中寻求一种“满意”的策略，可以保证列车安全、正点、运行过程舒适的同时，能耗和运行时间达到铁路运营部门和旅客都可以接受的程度，为达到这一目标，国内外铁路科技工作者进行了广泛深入的研究。本文就是以粒子群算法解决多目标列车运行过程优化问题，提出相应的改进优化算法。粒子群优化是一种新兴的基于群体智能的启发式全局搜索算法，粒子群优化算法通过粒子间的竞争和协作以实现在复杂搜索空间中寻找全局最优点。它具有易理解、易实现、全局搜索能力强等优点，所以设计起来更方便。为客观地描述列车的运行过程，建立了列车运行过程的多目标优化模型，并用粒子群算法求解该模型。针对多目标粒子群优化算法的不足，提出了相应的改进措施。仿真结果表明，提出的优化列车运行过程的改进多目标粒子群优化算法可以在一次运行过程中获得多组列车操纵控制策略，清晰地显示出各性能指标随着控制策略变化的趋势，控制序列转换次数大大降低，每组控制策略都可以在能耗、运行时间和停靠准确性之间获得很好的折衷效果，可以根据列车运行状况选择恰当的策略控制列车，以获得预期的结果。论文中围绕粒子群优化算法的原理、特点、参数设置等方面进行全面综述 。在MATLAB中对列车运行过程进行了优化，通过对优化结果的重现和比较，从而找到最合理的优化值。10530
关键词： 列车运行；多目标；优化；粒子群算法
Optimization of Train Operation Strategy Based on PSO
Abstract: This paper studies train operation models and saving energy operation optimization methods in railway traffic. Train operation process optimization is to find a "satisfactory" strategy from manipulation method, which can ensure the safety of the train, punctual, comfortable during operation process, in the meanwhile, energy consumption and operating time can reach the acceptable level of the railway operations departments and passengers. To achieve this goal, the domestic and international rail technology workers carried out extensive and in-depth research. This paper is the particle swarm algorithm to solve multi-objective train operation optimization problem and to propose appropriate improvements optimization algorithm. Particle swarm optimization is an emerging global based on swarm intelligence heuristic search algorithm, particle swarm optimization algorithm competition and collaboration between particles to achieve in complex search space to find the global optimum. It has the advantages of easy to understand, easy to implement, strong global search ability, so the design of it is more convenient. To reveal the essence of multiple objectives of train operation, a multi-objective model for train operation was established and solved by using the multi-objective optimization method. Improvement and keeping persity strategies were introduced to overcome the deficiencies of the existing MOPSO (multi-objective particle swarm optimization) algorithms. Simulation results show that the improved MOPSO algorithm can generate more than one train control strategy during a time running simultaneously, display changes in performance indices with the control strategies and decrease the shifting number of control serials sharply. Furthermore, fine tradeoff among energy cost, running time and stopping at adequate point can be obtained. As a result, the strategy suited to the train running can be selected to get an anticipated result. This paper gives a comprehensive description of the particle swam optimization principles, characteristics and parameters settings. The train operation process is optimized in MATLAB, and then to find the most reasonable optimization value by comparing and reproducing optimization results. MATLAB基于粒子群的列车运行过程优化:http://www.youerw.com/tongxin/lunwen_9618.html