Agent-Based Modelling And Simulation On Urban Green Space Planning Analysis For Disaster Sheltering

After decades of high-speed urbanization, China’s cities face a growing number of urban issues, urban planners have begun to reflect on the shortcomings of traditional urban planning methods, and to explore ways to improve urban planning techniques. Computer simulation models are effective techniques of quantitative analysis. As the current hot topic in the field of artificial intelligence and sociology, Agent-based model is getting more and more attention from urban planning scholars. Natural disasters occured in China frequently in recent years. The urban green space has an increasingly important role for its disaster sheltering function but there is no related planning standard laws and regulations, which lead to the deficiency of technical support. Based on the above the green space planning analysis was chosen as a starting point of Agent-based modeling research to explore its application in urban planning quantitative analysis.This research puts forward the basic ideas of the construction of green space planning analysis model for disaster sheltering and the quantitative parameters on the basis of the theories of green space system planning, comprehensive disaster prevention planning and the Agent-based modeling. After a comparative analysis, Repast S was chosen as the Agent-based modeling software, through which the model of green space planning for disaster sheltering was built. Then the model was tested by two planning cases of different scale. As to the data acquisition, the basic operation of the model was in accord with expectations, which could be used for the analysis and evaluation of green space planning for disaster sheltering and provide useful ideas for a similar planning analysis. The following is the main achievements of this paper:(1) This article discusses the application features, application conditions and possible application directionof Agent-based modeling in urban planning, clarified the definition of green space planning for disaster sheltering in their positioning in the planning system, summarized the specific process, and pointed out the defects of traditional planning methods. On the basis of the agent-based modeling principles and planning theories above mentioned a theoretical framework was put forward combined with quantitative analysis.(2) Through the study of software Repast S’s applications in urban planning, the CAD data import and integration methods with GIS. The key technologies of the application data conversion, Agent’s movement algorithm selection and the formulation of Agent’s action rules were achived.(3) Through the data tests on both large and medium scale planning projects, this Agent-based model had been verified effective. The workflow of data importing and model commissioning result in line with expectations, and this Agent model in planning analysis simulation was validated.