关键词: 开采沉陷, 概率积分法, 遗传算法, 支持向量机, 数字高程模型

Abstract: Traditional mining subsidence monitoring method with the shortcomings of time-consuming and low precision,which can not be used to predict the development trends of mining subsidence.Taking Pangushan tungsten mining area of Jiangxi Province as the study example,a new mining subsidence prediction method based on GA-SVM algorithm is proposed by the combination of GA (genetic algorithm) and SVM (support vector machine) algorithm.Firstly,the SVM algorithm is optimized by using GA,the operations of selection,mutate and crossover of SVM are cone by adopting GA;then,the mining subsidence prediction parameters of probability integral method based on GA-SVM algorithm,and the mining subsidence of the mining area is predicted.The study results show that the error between the mining subsidence prediction data of GA-SVM method and the actual measured data is lower than 5%,the DEM (digital elevation model) established by the mining subsidence prediction data by using GA-SVM algorithm is basically consistence to the one established by the actual measured data,which further indicated that the prediction precise of GA-SVM algorithm proposed in this paper is good.

Key words: Mining subsidence, Probability integral method, Genetic algorithm, Support vector machine, Digital elevation model