集成多智能体和复杂网络技术的城市扩展模拟研究

With China’s rapid social & economic development and accelerating urbanization, research on sustainable urban development is in the ascendant. Modeling / simulation is one of the effective ways to study urban growth phenomenon. This project starts with urban expansion analysis and hierarchical representation of transport infrastructure network. The travel and location choice models, which are based on multi-agent simulation technology, combine dynamic transport infrastructure network to improve the land-use changes sub-module. The transport network evolving model, which is based on the theory of complex networks, combines multi-agent simulation to improve the transport infrastructure network growth sub-module. These two interactive sub-modules are integrated to construct the final urban expansion model. The proposed urban model is applied in the study area and should be validated and evaluated. Scenarios-based simulation is set to predict the future urban development in order to support the management for planners and policy makers..The project is expected to: theoretically, propose a novel urban growth model based on multi-agent simulation and analysis and modeling of complex networks, integrating land-use changes and urban transport infrastructure network growth. The project not only promotes the development of urban modeling theory, but also enriches the GIS network analysis functions; algorithmically, design an efficient network topology construction algorithm and a fast the shortest path algorithm; practically, simulate and predict urban expansion in study area to significantly provide decision support for urban planning and management.

随着我国社会经济的快速发展和城市化进程的加速，对城市可持续发展的研究方兴未艾，而建模和模拟正是有效的研究途径之一。本课题在分析城市扩展机制、构建交通网分层表达的基础上，基于多智能体技术的出行和选址模型，结合动态交通网络，以改进城市土地变化子模块；基于复杂网络分析建模理论的交通网演化模型，结合城市主体的模拟，以改进交通网络增长子模块。由此提出一个集成这两个循环互动子模块的城市扩展模型。在特定研究区应用和检验该集成模型，并通过多情景模拟来预测研究区未来的发展以便于调控和管理。. 本课题预期在理论上，提出基于多智能体模拟和复杂网络分析建模的、集成土地利用变化和交通网络增长的城市扩展模型，推动城市建模理论的发展，也丰富GIS网络分析功能；在算法上，设计高效的网络的拓扑构建算法和最短路径算法；在应用上，模拟和预测研究区的城市扩展，为管理和调控提供决策支持，具有重要的实践意义。

针对中长期城市扩张模拟建模需求，本课题结合路网增长和土地利用变化两大相互影响的模块来研究城市扩张建模。路网增长模拟主要通过分析南京市路网演化情况，采用复杂网络技术，结合居民出行需求来建立增长模型；土地利用变化模拟主要采用基于路网约束的多智能体模拟技术，将常规的二维平面的分析方法拓展到路网约束空间中，使得智能体在出行模型和选址模型中的模拟过程中更符合实际情况。以南京市城市扩张2000-2010年为例，分别设置路网增长和土地利用变化交互关系强弱两种场景，模拟结果表明，强交互模拟结果更优，可解释性更强。本课题的研究结果强调了城市扩张建模不能忽视路网增长模块以及其与土地利用变化模块的交互作用。