Urban growth, urban sprawl if uncoordinated and dispersed, can be considered one of the most important policy agendas in modern urban regions. While no single policy option or remedy exists, understanding the urban growth system is the first step towards sustainable urban growth futures. Spatially explicit and dynamic urban growth models provide valuable simulations that encapsulate essential knowledge in planning and policy making such as how and where urban growth can occur and what the driving forces of such changes are. Over the past two decades, cellular automata (CA) models have proven to be an effective modelling approach to the study of complex urban growth systems. More recently Agent Based Modelling (ABM) has developed to yield a useful framework for understanding complex urban systems and this provides an arena for exploring the possible outcome states of various policy actions. Yet most research efforts of this sort adopt physical and heuristic approaches which tend to neglect socio-economic dynamics which is critical in shaping urban form and its transformation. This thesis aims to develop an agent based urban simulation model which has a more rigid theoretical explanation of agent behaviour than most such models hitherto. However, before developing such an agent based model, this study first conducted a series of experimental simulations with two well-known generic CA based urban models, SLEUTH and Metronamica, in order to better understand the complexity of designing and applying this class of urban models. Although CA and ABM are two distinctive modelling approaches, they share certain fundamentals concerning the complexity of systems and thus the empirical simulations with widely used CA models provide useful insights for the development of a new dedicated agent based urban growth model. For this purpose, each CA model is calibrated to the study area of the Seoul Metropolitan Area, Korea. The research then moves towards developing an agent based model based on microeconomic foundations. Utility maximising residential location choices made by households are modelled as the main impetus for urban growth through agglomeration and sprawl. Furthermore, based on such urban dynamics, alternative planning policy options such as greenbelts and public transportation are simulated so that their impacts can be clarified and assessed. In this way, the model is also able to examine how planning policies alter the economic utility of households and redirect market-led urban development. These results confirm the unique value of such modelling approaches. Yet, new research challenges such as the estimation of model parameters and the use of such models in planning support continue to dominate this field and in conclusion, we identify future research directions which build on these challenges
展开▼
