基于连续近似方法的公交网络与站点用地一体化战略规划

With the rapidly changing urban forms and growing congestion in Chinese cities, the traditional data-hungry and empirical methods for transit network planning have been proven to be time-consuming, expensive, inflexible and inadaptable, and often produce suboptimal solutions due to the mathematical complexity. In light of this, we propose in this project a parsimonious and computationally efficient method for optimizing urban transit networks that accounts for: i) the interactions between land use and the transit system; and ii) changing urban form and demand pattern. The proposed method will build upon continuum approximations (CAs) that represent transit demand and system characteristics as continuous functions or variables (e.g., line and stop spacings or densities and service frequency), and integrate models for station-area land use, residents/employers’ location choice, transit demand elasticity, and patrons’ budget constraints. We will explore the optimal decisions for three parties (the land-use developer, the transit agency, and the transit patrons) for various transit technologies, network structures, and land use scales. Specifically, we will investigate the following three research questions: i) optimal designs of multiple types of transit networks considering land-use response; ii) integrated planning of transit systems (especially the choice of transit technology) and station-area land use; and iii) scalable transit network design under dynamic socioeconomic conditions. Our models will be solved by exploring their analytical properties and using optimization tools. Case studies will be conducted for a number of Chinese cities (e.g., Chengdu, Beijing, and Hong Kong) with various operating environments. Results will be formulated into practical guidelines for medium- and long-term transit network planning as well as suggestions for transit metropolis and transit-oriented development.

在我国快速动态转变的城市环境下，传统数据依赖型公交规划模型建模成本高、周期长，其离散结构的数学问题求解困难、难以确保方案最优化。本项目拟采用连续近似建模方法将公交需求和系统特征（线路和站点间距或密度、服务频率等）描述为连续函数或变量，构造具有高效计算优势的公交网络设计解析模型，结合站点用地规划反映公交系统与城市用地间相互作用关系，探索决策指标（公交技术、线网结构和用地开发规模等）的全局最优解或相对关系。考虑用地开发者、公交企业和用户三类决策者，引入站点用地开发与居民/企业选址需求联合模型、弹性公交需求模型以及用户支出总预算约束，改写系统最优目标函数分别研究1）考虑用地反馈的多类公交线网设计；2）公交技术与站点用地一体化规划；3）动态社会经济条件下可扩展公交线网设计。采用解析推导和最优化工具求解上述问题，应用于我国多样化城市条件，总结指导中长期公交网络规划，也为公交引导城市发展提供建议。

本项目围绕公共交通线网设计中的关键科学问题，系统研究了基于连续体近似方法的公交线网设计优化问题，将公交需求和系统特征描述为连续变量或函数，构造了具有高效计算优势的解析数学模型，揭示了优化设计中关键决策变量与外部社会经济参数之间的规律性关系。相较传统基于离散模型的方法，本项目提出的连续近似模型，大幅提升了求解效率与质量，降低了模型对数据的依赖性。.具体研究课题覆盖了多模式网格型公交网络、共享站点的共享汽车服务、考虑环境影响的电动巴士系统、干线-支线公交走廊和网络、常规公交与共享单车混合系统以及需求响应式接驳公交系统。在土地与公交关系方面，研究了多中心走廊公交系统与居民选址联合优化设计，以及考虑不确定性需求增长的多阶段公交线网设计。运营方案方面研究了跨站停车、普速-快速等方案。本项目实际案例研究涉及了成都、重庆、深圳、雅安，华盛顿特区等城市，展示了所构建模型的实用性。