轨道交通出行需求与客流预测是涉及复杂人-地关系的非线性地理学问题。对城市建成环境与居民出行决策的复杂机理进行微观建模，进而构建耦合人-地关系的轨道交通出行ABM模型具有重要的理论意义。然而现有模型在居民与建成环境的精准表达与精细建模方面存在明显不足。本项目拟借助多源时空大数据对建成环境与居民进行精细识别，利用数据挖掘技术与ABM建模方法，建立精细化的轨道交通出行模拟模型。研究内容包括：（1）融合建筑物和大数据精细识别与挖掘建成环境与居民多维特征；（2）基于FP-tree算法挖掘居民与建成环境复杂时空关联规则，探讨居民活动与轨道交通出行行为的影响机理；（3）构建耦合居民与建成环境关系的轨道交通出行时空模拟的ABM精细化模型；（4）基于空间政策与规划情景模拟轨道交通出行行为，为新型交通服务模式与规划方案的影响评估提供科学、有效的决策支持。

Urban rail transit travel demand and passengers flow forecasting is a non-linear geographical problem involving complex man-land relationship. Exploring complex mechanism between urban built-environment and travel decisions of residents at a micro scale, then building a ABM model for rail transit travel behavior has important theoretical significance. However, there are still some insufficiencies in existing models. They cannot accurately delineate and model residents and built-environment. Therefore, in this project we will use multi-sourced spatio-temporal big data to accurately delineate urban built-environment and residents. We will also establish fine-scale simulation model for rail transit travel behavior by using data mining techniques and ABM method. The contents of our project include: (1) Accurately delineating multi-dimensional characteristics of built-environment and residents based on the integration of buildings data and big data; (2) Mining the spatio-temporal association rules between residents and built-environment by FP-tree algorithm, and further exploring the influence mechanisms of residents' activities and rail transit travel behavior; (3) Building fine-scale simulation model for rail transit travel, which couples the relationship between residents and built-environment. (4) Simulating the spatio-temporal rail transit travel behavior based on the transportation policy and urban planning scenario, which can provide scientific and effective decision support for impact assessment of new transit service patterns and planning programs.