Land Use Planning Methods Structure Design And Applied Research Based On Artificial Neural Network

The revision and implementation of general land use planning is a systems engineering, and the final planning project is seriously affected by nature, society, economy and so on. Social problems are too complex and the future is uncertain during the rapid developing period of economy. It has caused the land use planning techniques transferred from linear planning in the early to multi-objective linear planning, and even nonlinear directions later. The forms of planning also have been transferred from static planning directions to dynamic planning directions. The original too rigid planning has becoming taking on certain flexibility. The traditional planning methods were too difficult to meet the demands of the complicated planning process. This required land use planning to consistent with the objective reality by using new theories and new methods in the process of land use phnning revision and execution management, dealing well with the relationship between flexibility and rigidity, static and dynamic in order to make the land use planning more operable. Therefore, this article takes the micro-level general land use planning revision as a breakthrough point in the perspective of model methodology, integrated the artificial neural network methods and general land use planning revision technology. It took Xinyu city of Jiangxi Province as a example, and established the neural network model of the land use planning in the special study of the new round general land use planning of Xinyu city. It also discussed questions appeared in the planning practice based on the artificial neural networks model on the foundation of land use planning model method theoretical analysis. In sum, the contents of this paper can be drawn mainly as follows:1. proposed the new thoughts of land-use planning model methodologyLand use planning is a systems engineering, and the model-based approach is its main method. The innovative application of Land planning model-based approach is the main body of planning techniques methods innovations. The existing traditional planning model methods were limited in emphasizing the linear or the deterministic relations between planning factors. They are methodologies based on the inference and the gradually style. It is also difficult to deal with the nonlinear, recessive, indeterminacy functional relation in the land use planning, and satisfy, the dynamic and elastic demand of land use planning. The request of the complex planning process is also found to be difficult to adapt. Thus it is indicated that certain flaws and insufficient exist in the traditional plan model methods. This article discussed the mathematical models and their application is a rational choice of land use planning system modeling theoretically in the foundation of scientifically recognized the land use planning methods and conducted the research into the existing planning model methods. Two new thoughts were put forward in this paper:complementing the main mathematical models of land use planning; introducing the new methods of nonlinear data processing -- the artificial neural network models.2. choose the land use planning models and the algorithm based on artificial neural networkOn the basis of introducing the artificial neural network theory and its principles, this paper has emphatically analyzed the BP network models, the radial basis function (RBF) network models and the linear neural network models, which are the most important forward feeds neural network models. Then it discussed some problems of these three kind models such as the network structures, the algorithm mathematics description, the study parameter setting... It also elaborated the realization process of the artificial neural network models based on MATLAB software by using the high-level algorithms of high-performance numerical computing and neural network toolboxes of MATLAB software. GUI realization methods and the M language programming realization methods based on MATLAB software were also elaborated in detail in this paper.3. the theoretical design and application of land use planning forecast models based on artificial neural networkForecast analysis is an important part of land use planning. It is the premise and the key of land use planning and a difficult problem of land use planning study. Whether the land use planning will be successful depends on it. Because the complexity of the data, the nonlinear and uncertainty of the land-use planning forecast study, traditional forecasting methods are often ineffective. Thus, this paper proposed a land use planning forecast method based on the artificial neural networks model on the basis of analyzing the characteristics and problems of the tradition forecast method modeling. It has also analyzed the feasibility of establishing two types of neural network land forecast models: the time series and the regression analysis. Took the time series data type as an example, it has carried on detail technical design for land forecast model based on the artificial neural networks, including the network model choice, the network structure design, the sample construction, the parameter setting and so on. Took the total history population forecast of Xinyu city and Fuzhou city of Jiangxi Province as a confirmation example, it established a population forecast linear neural network model and carried on a simulation forecast in order to confirm the forecast ability artificial of the neural networks model, and then analyzed the similarities and differences between the artificial neural networks forecast model and the traditional forecast methods.4. the theoretical design and application of the land use optimization models based on artificial neural networkThe ultimate goal of Land use planning is to produce one or several planning projects that can guide reasonable use of future land. The future land use structure is the core content of planning projects. The adjustment and optimization process of land structural is going with the adjustment of the planning projects. So the research on the adjustment and optimization of land structural is the core of land use planning. At first, this study analyzed the feasibility and the necessity of the artificial neural network model combined with the optimization and adjustment of land use structure on the basis of traditional theories and methods of the optimization and the adjustment of land use. It took three levels forward feeds BP network as an example, and carried on the technical design based on the artificial nerve network for land use structure optimization model. Secondly, from aspects the system description, model structure, logic analysis, inference and explanation and so on, it compared and the identified the artificial neural network models and the system dynamics SD models in a theoretical level-- two effective computer dynamics simulation methods.Finally, it applied the land use structure optimization model that based on artificial neural network to new round of general land use planning revision research of Xinyu city, Jiangxi Province. The feasibility and the efficiency of the model were testified by examples in order to improve the quality of the land use planning of Xinyu city.5. the theoretical design and application of land use planning evaluation models based on artificial neural networkIn the each steps of land use planning compilation, approval and execution, evaluation is an extremely important content. Tradition land use planning evaluation methods have many insufficiencies, whether it can be successful lies on choosing a good evaluation method, determining a evaluation target system, choosing appropriate target weight and so on. Therefore, this research technically designed the land use planning evaluation model based on the artificial neural networks on the basis of analyzing and comparing the insufficiencies of tradition evaluation methods, and proposed four effective methods of constructing samples in the planning project evaluation model. It also designed different types of utility functions of the land use planning program evaluation standardization. The article took Xinyu city of Jiangxi Province as an example, and used the RBF network structure to evaluate the land use planning projects and then choose the best. It has constructed a real diagnosis model of land use planning project evaluation of Xinyu city.The result indicated that the land planning evaluation methods that based on the artificial neural networks model have the characteristics such as better evaluation effects, stronger objectivity, and more flexible compared with the tradition evaluation methods. It also pointed out that the question that may effected the artificial neural networks evaluation model was that the number of study samples of the planning program evaluation was too small.