Research On Land Reduction Simulation And Performance Evaluation Based On Cellular Automata And Neural Network

For the first time,this paper uses cellular automaton model based the neural network to simulate the construction land reduction in Shanghai,and analyzes the influencing factors behind the construction land reduction to reduce the construction land in Shanghai.Quantitative performance is assessed and recommendations and recommendations are made based on different influencing factors.At present,there are many simulations and predictions of urban land use evolution through the cellular automaton model.The previous models are usually applied to the simulation of the conversion of non-urban land into urban land,but the evolution of various land use types.The simulation is far more complicated than the previous simulation of urban evolution.The simulation of the evolution of various land use types requires the determination of corresponding environmental parameters and spatial variables in advance,and the determination of the relevant parameters and variables of cellular automata is current.The difficulty to be solved.In response to this problem,this paper will use the self-learning characteristics of BP neural network,which can effectively simulate the change of construction land in Shanghai,and greatly shorten the time required to determine the conversion rules of cellular automata model.Excavation of the change law of land types in the past can predict the land use situation of Shanghai in 2017,and compare the forecast results with the actual land layout of Shanghai in 2017 to verify the model.The simulation results are less accurate,and the relevant parameters of the neural network are changed and debugged.When the accuracy of the simulation results meets the requirements,the land use situation of Shanghai is predicted by the excavation evolution law,and the predicted construction land and the construction land of the previous year are processed in Arcgis,and the Shanghai 2017 is obtained.By 2020,the reduction of construction land per square kilometer will be analyzed,and the factors affecting the construction land will be analyzed.Different influencing factors will be preprocessed and introduced into the cellular automaton model,and different influencing factors can be obtained.The impact on the amount of construction land reduction.According to the simulation results,the performance of the reduction of construction land in Shanghai is evaluated and the actual solution is proposed based on the analysis of the trend of the reduction of construction land per unit area according to different influencing factors,which provides a reference plan for the efficient implementation of construction land reduction in Shanghai.This article builds the model from the following sections:1.Establish a land cell sub-database to provide a data foundation for model building.Through Arcgis,the land use vector data of Shanghai is rasterized,and the grid size is determined according to the operation speed and accuracy requirements.In order to train the land use data through the neural network,the formed raster data needs to be ASCII by Arcgis.Processing,generating a matrix form database that can be processed by Matlab,that is,a land cell sub-database,optimizing the land cell subdatabase,dividing the land cell sub-database into a training sub-database,verifying the sub-database and the prediction sub-database.2.Using BP neural network to mine the law of land use evolution.The Matlab module is compiled and invoked by Matlab to train the data of t time and t+1 time in the training sub-database in the above-mentioned land cell sub-database,that is,the land use data of Shanghai in 2012 and 2017 are carried out.training.Among them,the land use data of Shanghai in 2012 is the input data of the neural network.In 2017,the land use data of Shanghai is the target data of the neural network.Through the selflearning characteristics of the neural network,the law of land use change in Shanghai is reversed,that is,the conversion rules of land cells are determined.3.Verification and prediction of land use change in Shanghai.Based on the land use law of the above-mentioned inversion,the land use distribution data of Shanghai in 2012 will be predicted,and the land use distribution of Shanghai in 2017 will be predicted.The land use change in Shanghai in 2017 will be predicted and the actual 2017 Shanghai.The land use distribution is compared and the predictions are tested.When the accuracy of the model meets the requirements,the land use data of Shanghai in 2017 will be used to predict the land use change in Shanghai in 2022.4.Analyze the impact of different influencing factors on the reduction of construction land in Shanghai.Through Arcgis,the construction land in Shanghai in 2017 and the predicted construction land in Shanghai in 2022 will be layered and differentiated to form a grid.Statistics on the reduction of construction land in the grid will be obtained.The reduction of construction land in the city from 2017 to 2022.Through in-depth analysis of the influencing factors of the construction land reduction in Shanghai,and the raw data of different influencing factors are pre-processed,the model is introduced as an input factor,and after adding different influencing factors,the 2017 Forecasting the changes in construction land reduction in Shanghai during 2022,and combining the previous construction land reduction in Shanghai to reduce the actual work situation,evaluate the performance of the reduction work,and propose corresponding problems for the existing problems.Suggestions to provide reference for the effective implementation of the reduction work.The area of construction land predicted by CA-BP model is higher than that planned by the government.It is necessary to further strengthen the reduction of inefficient construction land to achieve land saving.