交通运输工程与信息学报

确定城市轨道交通有效路径是客流预测、客流清分等运营管理工作的重要基础,本文在分析乘客对换乘次数及出行成本容忍度的基础上,设计了基于广度优先算法的双向BFS(breadth-first search)算法。考虑拓扑网络中的换乘站、非换乘站、环线与虚拟换乘弧的特性,对轨道交通站点、线路、区间进行编号,并根据乘客换乘次数不超过3次的行为特征,借助相邻换乘站确定两条路径交汇的换乘站,然后通过引入广义相邻换乘站的概念筛选重复路径,标定广义费用函数计算出行成本。最终设计出从O、D点同时开展广度优先算法的双向BFS算法,该算法具有适用性强、速度快等优势。以成都地铁网络中茶店子站和磨子桥站这一对OD为例,在6条初始路径基础上经筛选最终获得2条有效路径,结果完备无冗余。

The effective route of urban rail transit is necessary for passenger flow forecast and classification. Based on the analysis of passengers' tolerance to transfer times and travel costs, a bidirectional breadth-first search(BFS) algorithm was designed in this study. We successively considered the characteristics of transfer and non-transfer stations, loop lines, and virtual transfer arcs in the topological network, and then completed the numbering of rail transit stations, lines, and sections. Based on the passengers' behavioral characteristics with no more than three transfers, the transfer station where the two paths met with the help of adjacent transfer stations was used to filter the repeated paths by introducing the concept of generalized adjacent transfer stations. Next, the generalized cost function was calibrated to calculate the trip cost. Finally, a bidirectional BFS algorithm that performed breadth-first algorithms from both points O and D was designed, and the algorithm had advantages of strong applicability and high speed. The Chadianzi and Moziqiao stations in the Chengdu subway network were selected as a pair of ODs for demonstration. Based on the six initial paths, two effective paths were finally determined after screening, and the results were complete with no redundancy.