基于旅客类别的列车服务网络客流分配

Train service network flow assignment based on passenger classification

针对列车服务网络客流分配模型复杂、求解困难的特点,采用旅客类别划分的简化策略,以旅客主体特性为分析参数,运用近邻传播算法对其进行聚类。在此基础上,构造列车开行方案形成的服务网络,提出不同类别旅客乘车方案弧段阻抗的计算方法,建立基于类别划分的旅客乘车方案概率选择模型,并设计相应的网络流量加载算法。以宝成线为实例,根据客流调查结果,进行旅客类别聚类划分与流量加载验证。研究结果表明:将铁路出行旅客划分为6个类别时,具有最好的聚类效果;在此基础上客流分配准确率在80%以上。基于旅客类别划分的流量分配方法能够得到准确的流量分配结果,从而为列车开行方案的调整提供合理依据。

According to the railway network multi-flow path optimization problem of passenger train plan. The simplifying method of passenger classification was firstly adopted, which takes individual feature of passengers as analyzing parameter. And affinity propagation algorithm is used for the cluster of passengers. On that basis, service network constructed by train plan is built, together with the network impedance of different passengers. Then the passenger flow assignment model of train service network reflecting passenger category was established, and relevant strengthening algorithm is designed. Taking Baoji-Chengdu Railway as an example, its cluster partitioning on passenger categories and traffic load verification was analyzed according to passenger survey results. Cluster analysis results show that when the railway travel passengers are divided into six categories, the clustering effect is the best, and on this basis, the passenger flow distribution also shows that the accuracy is above 80%. The flow distribution method based on the classification of passenger categories can get accurate flow distribution results, so as to provide a reasonable basis for the adjustment of train operation plan.