大数据支持下的精细土地利用与多模式交通网络协同优化

Traffic congestion is the “city disease” which is a common occurrence in our present cities. The reason for this is the traffic imbalance between supply and demand. The comprehensive planning of transportation and land use is the key to solve this problem since they are the sources of traffic supply and demand. Thus, exploring the complex mechanism between land use and transportation has important significance for their coupling simulation and collaborative optimization. However, at present the theories and methods in this area is still in initial stage. In this project, we use advanced technologies, such as data mining, geographical simulation and optimization, to establish the coupling simulation and collaborative optimization model for fine-scale land use and multi-modal traffic network by combining the increasing big data and small sample survey data. Concrete research contents include the following several aspects: 1) exploring the microscopic interaction mechanism between land use and transportation; 2) integrating the MAS simulation and artificial immune algorithm to construct the land use change simulation and multi-scale optimization under the guidance from public transportation; 3) setting up a prediction and optimization model of multimodal transportation passenger flows by using individual behavior simulation and ant colony algorithm; 4) building a spatial-temporal collaborative optimization model for land use and multi-modal traffic network. Through the fine simulation and collaborative optimization research, this project aims to provide scientific basis and effective supports for developing a scientific, prospective and harmonious planning of land use and transportation.

交通拥堵是当前我国城市普遍面临的“城市病”，其根源在于交通供需失衡。交通与土地利用是交通供给与需求的来源，两者的综合规划则是解决这一问题的关键。因此，探讨土地利用与交通之间的复杂机理，对二者进行耦合模拟与协同优化具有重要的意义。目前这方面的理论和技术方法尚处于初始阶段。本项目拟结合日渐丰富的大数据与小样本调查数据，利用数据挖掘、地理模拟与优化等先进技术，建立精细土地利用与多模式交通协同模拟与优化模型。具体研究内容包括：1）探讨精细土地利用与交通网络的微观互动机理；2）结合MAS模拟与人工免疫算法建立公共交通导向下土地利用演变模拟与多尺度优化模型；3）利用个体行为模拟与蚁群算法，建立多模式交通客流预测与优化模型；4）构建时空环境下土地利用与多模式交通的协同优化模型。本项目旨在通过对两者进行精细模拟与协同优化研究，为制定科学性、前瞻性和协调性的土地利用与交通综合规划提供理论方法与技术支持。

本项目结合日渐丰富的大数据与小样本调查数据，利用数据挖掘、地理模拟与优化等先进技术，在地理空间信息感知与快速识别、土地利用变化模拟和城市增长边界划定等多方面取得许多创新性的研究成果，包括：融合高分遥感影像和多源社交媒体数据的城市土地利用分类；基于 GEE 平台的广州市主城区不透水面时间序列提取；基于多源空间融合的建筑物尺度级城市人口空间分布精细制图；结合多源大数据识别建筑物功能；基于随机森林的多类城市内部土地利用变化模拟及驱动力因子分析；耦合人类活动与自然效应的未来土地利用变化情景模拟模型；耦合动态地块分裂和矢量元胞自动机的城市发展过程精细化模拟；基于梯度元胞自动机模型的中国城市动态模拟；基于地理分区与FLUS模型的城市扩张模拟与预警；基于FLUS模型和形态学方法划定多种情境下城市增长边界；耦合资源环境承载力评价与城镇空间适宜性评价的重庆市城镇开发边界划定研究。项目成果能够反映城市的扩张进程和复杂的土地利用发展过程，为国土空间规划提供有重要的技术支持。在项目期间共发表 SCI/SSCI论文29 篇，培养博士8名，硕士10名。