Spatial Layout Optimization Of Rural Residential Areas Based On Cellular Automata And Particle Swarm Optimization

The development of social economy and the upgrading of the industrial structure have deeply changed the urban-rural relationship. As an important part of the urban and rural construction land, rural residential areas are the major medium and conversion content of rural urbanization and intensification. Its spatial layout not only reflects the rural settlements relations of production and social culture, but also determines the functional configuration and overall efficiency of rural land use system. With the in-depth propulsion of new countryside construction and the rapid integration process of urban and rural areas, layout optimization of rural residential areas inevitably becomes the top priority for structure optimization of rural-urban land utilization. However, layout optimization of rural residential areas is a complex and nonlinear geographical process which multiple restricted and influenced by regional nature、society and economy. How to acquire effective knowledge and mode of layout optimization of rural residential areas from massive land use data to promote it more scientific and intelligent with the combination of geographical simulation system and intelligent optimization, which will have become an important branch of the research on structure optimization of land utilization.The solutions to spatial optimal search are always complex and important in the GIS applications. Multi-objective and constraint conditions make it necessary to find reasonable theory and method to solve these problems. There are two types of village distribution optimization:urbanization and intensification, they are orderly in process and simultaneously at the same time. Village urbanization integrates spatial data model and spatial process model based on grasping the process and laws of urban sprawl, it demonstrates the process of urban and rural amalgamation. The paper analyzes reciprocal mechanism of micro-spatial entity by bottom-up strategy, which involves in the thinking of multi-agent decision and game based on the analysis of urban extension mechanisms. The research integrates multi-agent system、cellular automata and GIS to explore and analyze the formation and evolution of urban land distribution, and to plan the distribution of settlements which will be swallowed by towns. Village intensification embodies land intension and resource distribution. The optimal target varies with the subjective and objective conditions, it attains Pareto optimality by stepwise optimization on the fundaments of reasonable goals. This paper attempts to introduce the particle swarm optimization and combine it with GIS to realize optimal goals of rural residential areas. It iterates and updates through searching optimum position of colony and individual based on numerical and spatial restriction.Previous research of rural residential areas placed extra emphasis on evaluating of spatial features and probing of evolution mechanism on its content, lacked integration of GIS to solve spatial decision problems in its method, limited to quantitative optimization and intensive evaluation in its model. The applications of geographical simulation system and intelligent optimization are almost vacant in the distribution optimization of rural residential areas. To solve these problems, this paper points out the rationality and necessity of applying cellular automata and particle swarm optimization to optimize of rural residential areas, puts forward the research idea of its optimization based on cellular automata and particle swarm optimization.The spatial distribution of village urbanization not only reflects the transition of land type, but also embodies the mechanism of geographical environment and behavior game of government, enterprise and household. The key of model construction is reflecting the coupling relationships of rural and urban system, natural and social system, spatial and temporal system. The paper has built up a series of converting rules and constructed a dynamic urban land expansion model to realize conversion and unity of spatial data structure. The model integrates social organization and spatial expansion advantages of multi-agents and cellular automata at the urban micro level, to analyze the interaction process and complex spatial decision of urban geographic space from an evolutionary and emergent perspective.The important core issue is to construct optimization model of village intensification based on multi-objective particle swarm optimization. Under the villatic distribution optimization theory guidance, the paper states the research contents of the optimization model in the following aspects:village suitability evaluation, multi-objective restriction, algorithm design. In addition, the research discusses two patterns of distribution optimization:central village guidance and primary village guidance, according to the terrain feature. Combining with numerical restriction, land restriction and spatial restriction, the paper establishes multi-objective optimization functions based on land compatibility, conversion accessibility and spatial compactedness to realize distribution optimization of rural residential areas.The paper takes Jiayu county as an example to explicate the rationality and feasibility of system structure, distribution model and optimization algorithm, selects Yuyue town (central city), Guanqiao town (hilly and industrial town) and Panjiawan town (plains and farm belt) as three different types to compare the spatio-temporal difference and internal mechanism of village dynamic evolution by using land use data and relevant statistic data of 2005 and 2010, introduces cellular automata, multi-agent and particle swarm optimization to establish layout model. The result shows feasibility and superiority of the research by comparing and analyzing to the distribution optimization of rural residential areas from 2010 to 2020.