时空双视角下基于多源数据的城市轨道交通网络配流评价方法与失效机理研究

城市轨道交通网络配流模型是城市轨道交通系统客流分布计算及票款清分清算的核心模型，当前已从模型构建和参数标定进入到了评价验证及改进优化的新阶段，而使用的主要还是一类基于人工客流调查的方法。本项目拟从城市轨道交通大规模网络化运营管理的实际需要出发，在空间与时间的双视角下解析网络配流有效性的准确内涵与评价标准；基于对AFC票卡、ATS行车、CCTV视频、WIFI嗅探、手机信令，以及客流调查等多源数据资源的综合使用与融合分析，构建多源数据驱动的城市轨道交通网络配流评价框架体系，在空间和时间两个维度上建立配流有效性的评价方法及模型；基于实证研究和仿真分析的方法研究现有网络配流模型存在的主要问题与影响因素，揭示网络配流模型计算的失效机理，为模型系统的持续改进及优化提供理论指导，丰富和完善城市轨道交通网络客流分布计算理论，不断提高城市轨道交通系统的现代化运营管理水平。

The assignment model is the core for calculating the passenger flow distribution and clearing fares on a rail transit network. At present, the rail transit assignment model is entering into the new stages of evaluation and improvement from the past stages of modeling and calibration, however, the currently-used approaches for evaluating assignment models are simple and mainly based on manual methods. In this study, we try to define what is an “effective” assignment model and how to evaluate it from the perspectives of both space and time, using multi-source data including automatic fare collection data, automatic train supervision data, closed-circuit television data, wireless fidelity detection data, mobile signaling data, and investigation data. The problems and influence factors in assigning passenger flows on network are investigated based on empirical study and simulation analysis, and then the failure mechanism of rail transit assignment models is revealed. This study effort will provide theoretical guidance for the continuous improvement and optimization of rail transit assignments, enriches the theory for calculating passenger flow distribution on a rail transit network, and consequently help to improve urban rail transit operation and management.