Multiscalar modeling of polycentric urban-regional systems: Economic agglomeration, scale dependency and agent interactions

This dissertation aims to study the causal relationship between the underlying processes of agglomeration economies and the formation of certain spatial structures at both intra- and inter-urban scales within the extent of megaregions. First, on the theoretical side, I develop a general framework to account for the interplay between market linkages and spatial costs across scales. The model system constructed from this framework allows us to study the evolution of intra- and inter-urban spatial structures in terms of monocentricity/polycentricity and agglomeration/dispersion. By examining the impacts of local spatial costs and interregional trade costs on the structural change at both scales, I find the interdependency of spatial structures across urban-regional scales. Second, I extend the theoretical models into a 2-D geographic model that can scale to real world applications. This geographic model is based on zonal geography connected by transportation networks. An agent-based approach is employed to model the discrete choice for locations and to approximate the equilibrium conditions as those in theoretical models. The simulations confirm that it is consistent with the theoretical models regarding the generated spatial structures. A demonstration application to the Carolinas megaregion is presented as a test bed for the geographic model. Three simulation scenarios presented provide insights to understand the observed pattern of urban-regional developments. In addition, high-performance computing technologies are leveraged to improve the computational performance of the geographic model.