| In recent years, the process of urbanization and industrialization continues to accelerate, and the issues come out such as excessive expansion of urban land, infrastructure duplication and urban development disorder. In order to make full use of our land resources, and make sure that all factors of production can be efficiently allocated in the level of region, we have paid attention to the planning of regional land use, continue to strengthen the preparation of the national and provincial planning of land use, accelerating to build regional planning law and related legal system, then we can promote the science and authority of regional policy, strengthening the role of regional planning strategy to guide the development of our country.To meet the needs of regional land use planning, scholars began to research the urban land development at the regional level so that develop regional land use planning for the national and make full use of land resources. Current research about the development of urban land in mostly the urban area, the main focus on an inner-city urban land development mechanisms, which means to build a unified set of land use conversion rules to simulate the entire study area of the urban land development. Due to natural resource endowments, economic and demographic background factors such as infrastructure, are very different from other cities within a region, all the city’s land use development rules and driving factors showed a significant difference in space. Therefore, when performed for the regional level analysis of the development of urban land, we need to improve the traditional simulation methods, urban land emphasis on intra-regional differences in the evolution of the rules of space to enhance the regional acuracy of urban land use simulation.This paper has theoretical significance and application value to the simulation of urban development at the regional level, and the development of regional land-use planning.In this paper, we analyze the limitations of the traditional CA model at the regional level, we combined the method of case-based reasoning(Case-Based Reasoning referred CBR) and cellular automata(Cellular Automaton referred CA) to create the Case-based reasoning and Cellular automata model(CBR-CA) on the regional level to simulate the development of urban land. The main contents include:(1)Try to make and organize the factors affecting land use types such as feature attributes, and spatial relationships in the form of geographical cases;(2)Sample cases to build the geographical cases database in the regional level, and use the geographical cases database to predict the future development of urban land in the regional level;(3)Combine the result of the case-based reasoning to CA model to simulate the spatial pattern of future urban land.(4)As a study area in Hunan Province, study the difference of Hunan Province urban land use rules, and use case-based reasoning-CA model to simulate the urban land development in 1990 and 2010 of Hunan Province.(5)Analysis the accuracy of CBR-CA model simulation to verify the applicability of CBR-CA model for the simulation of the development of urban land in region.Through the analysis and discussion, we have the following conclusions:(1) The use of CBR-CA model based on the regional level to simulate the development of urban land is reasonable. The traditional CA model simulation of urban land use development by building specific transformation rules, CBR-CA model use geographic cases to express implied evolution of the land within the region spatial differences and work structure eliminates the cumbersome and complex land conversion rules of urban land area; in addition, compared to traditional CA model, CBR-CA model can affect the orientation of the geographic units, spatial distance and land type topologies such as the evolution of urban land spatial relationships applied to simulate the development of urban land, which is more comprehensive in-depth expression, refined the roles of urban land, more realistically simulate the development of urban land, which is a complex geographic phenomena.(2) This paper studied the different of development roles in Hunan Province, found that the effects of different cities in different regions of the driving factors of land, land use and transformation of spatial differences are significant, but there is a certain spatial correlation in each city driving factors and evolution. Study urban land spatial variation characteristics will help to guide the sampling CBR-CA model in a reasonable construction of the case database, also help to study the further mechanism and characteristics of urbanization.(3) Use CBR-CA model to simulate the development of urban land in Hunan Province, and evaluate the accuracy of the simulation results. Experimental results show that, the result of CBR-CA is better than the result of traditional CA model. Among them, from a province-wide perspective, the accuracy of result based on the logistic simulation is 74.2%, and the accuracy of result based on case-based reasoning is 82.4%; both from the province, regional and county level, the simulation results based on CBR-CA are better than the logistic. |
PDF Full Text Request