应用神经网络优化高速铁路动车组周转的研究  

Study on the optimization of high-speed Une train-set scheduling using neural network

作　　者：陈华群[1] 唐协[1] 

机构地区：[1]西南交通大学,成都610031

出　　处：《中国安全生产科学技术》2006年第2期46-49,共4页Journal of Safety Science and Technology

摘　　要：结合当前国内外对高速铁路动车组运用的相关研究,给出了动车组周转这一特大型组合优化问题的数学描述并建立神经网络模型,提出了基于分层聚类思想与模拟退火算法相结合的解决方案,降低了算法的时间复杂度,达到了减少车底在车站的停留时间,提高了车底利用效率,为优化我国在建和拟建的高速铁路、城际客运专线的动车组周转及计算机自动编制车底运用计划提供理论支持,并结合实际客运专线运用计算机模拟进行检算,证实了算法的可行性、实用性。Based on the domestic and overseas correlative research of high-speed train-set schedule, the paper introduces mathematic description and neural network model of the high-speed train-set scheduling. Then the paper uses neural network to solve the super combinatorial optimization problem and analyses the theory course of using Hopfield neural network. Then the discourse brings forward sorting and clustering idea combining with simulated annealing to ameliorate times complexity. This method reduces the stay time and improves the efficiency of train-set schedule. In a word, the study provides the theory of optimizing and using computer to workout the train-set scheduling of the high-speed or special passenger railway. At last the article combines the actual special passenger line and use computer simulation to approve the feasibility and practicality of the study.

关 键 词：高速铁路 动车组周转 HOPFIELD神经网络 分层聚类 模拟退火 

分 类 号：X951[环境科学与工程—安全科学]