Study On City Growth Spatial Form Based On Constrained Logistic-CA Model

Understanding the principle of spatial form evolvement in the process of city growth is remarkably meaningful for city planning, especially for developing China which is in the phase of rapid urbanization. The self-organization feature of "bottom-top" in city growth conforms well to the self-organizing nature of Cellular Automata (CA). So, it is a road worth exploring to establish a model based on CA.In this thesis, constrained logistic-CA model is established on the base of previous study. This model adopts the logistic regression method to mining the effects of several location feature variables to cells status conversion, consequently enhances the facticity and scientificity of the model, avoids the subjectivity of manual configuration of influencing factors' weight. The model also introduces the constraint effect and the random effect to get over the low facticity problem of standard CA in geography research. With respect to the neighbor effect which is the kernel of CA, a new controllable pattern is introduced in constrained logistic-CA model. Different spatial forms can be simulated by using different neighbor factors.After the implementation of constrained logistic-CA model in ArcGIS Model Builder, this thesis takes the city zone of Changsha and the vicinity as research region to carry out experiments with the model. Firstly, the processes of city growth in two periods of 1993-1998 and 1998-2006 in the research region are simulated with precision of 87.92% and 76.51% respectively. Then the predictive simulation of city growth from 2006 to 2022 in the research region is done with two different groups of model parameters.To analyze the spatial form of city in the research region and the results of model simulation more scientifically, this thesis introduces relative method from landscape ecology, extracts a group of metrics from landscape metrics to evaluate city spatial form and analyzes the spatial form change of real and simulated Changsha quantitatively. The analysis shows that constrained logistic-CA model can simulate the changing trend of the city spatial form in the research region, but the ability to simulate the absolute value of metrics is not enough. In the end, the thesis concludes that Changsha will mainly take spreading growth pattern in the future through analyzing two simulation results of period 2006-2022.