Abstract:

To quantify the impact of passengers' transfer on their travel route selections, this study added virtual transfer stations to the single-layer network, and developed a three-layer multi-standard rail transit topology network model without hidden connections. The connectivity coefficient between networks was calculated based on the time and transfer node connectivity. A function was established to describe the passenger's familiarity to the line network. In this function, the Dijkstra method was used to search for the K short path between the start and end points of the model, and the time was determined in consideration of passengers' travel plans and the complexity of the line network from the passengers' perspectives. A generalized travel cost function was also established to reflect the factors affecting the passenger path selection. The Logit model was used to calculate the selection probability for each route. The passenger flow distribution algorithm was designed to solve the model and then complete the passenger flow distribution to the multi-standard rail transit network. The multi-standard rail transit network simulation model was based on the Chengdu Metro, Chengguan, Chenggui High-speed Railway lines, etc. The passenger flow distribution examples analysis show that the passenger flow distribution results are basically consistent with the actual data, which confirms the authenticity and effectiveness of the passenger flow distribution algorithm.

Key words: railway transportation, passenger flow distribution, travel route selection, multi- standard rail transit, line network familiarity, Logit model

摘要：

为量化换乘对乘客出行路径选择的影响程度，在单层网络中添加虚拟换乘站，构建无隐性连接的三层多制式轨道交通拓扑网络模型。基于时间、换乘节点衔接性，计算线网间衔接性系数；利用Dijkstra法搜索模型各起讫点间的 K 短路径，以乘客感受到的线网复杂度及乘客出行计划确定时间，建立乘客对线网的熟悉度函数；根据乘客路径选择影响因素构建广义出行费用，利 用Logit函数对每条路径的选择概率进行计算；最后设计客流分配算法进行求解，实现对多制式轨道交通网络的客流分配。以成都地铁、成灌、成贵高铁等线路建立多制式轨道交通网络仿真模型，对其客流分配实例分析表明，客流的分配结果与实际数据基本吻合，证实了客流分配算法的真实有效性。

关键词: 铁路运输, 客流分配, 出行路径选择, 多制式轨道交通, 线网熟悉度, Logit模型