关键词: 量子神经网络, 二次多项式拟合, 变形监测, BP神经网络, 残差, 内外符合精度

Abstract: The surface deformation monitoring of mining area is influenced by the differences of geological structure conditions,mining scale and mining methods.The classical surface deformation monitoring methods of mining area with the defects of process trival,heavy workload and low accuracy,so,the surface deformation monitoring method of mining area based on quantum neural network fitting method is proposed.The coordinates (x,y) of monitoring points in mining area are taken as the input layer neurons of quantum neural network,the abnormal height (ξ) of the monitoring points in mining area are taken as the output layer neurons of quantum neural network,optimal solutions are obtained by multiple iterations of quantum neural network.Based on the GPS monitoring data of a mining area,the quadratic polynomial fitting method,BP neural network fitling method and the quantum neural network fitting method proposed in this paper are adopted to conduct contrast experiment,the internal precision and external precision are taken as the evaluation standard of the fitting precision of the above three methods.The experimental results show that the internal precision and external precision of quantum neural network fitting method proposed in this paper are higher than the quadratic polynomial fitting method and BP neural network filling method,the residual errors of the quantum neural network fitting method proposed in this paper is lower than the quadratic polynomial fitting method and BP neural network filling method,the experimental results further show that the quantum neural network fitting method proposed in this paper has reference for the improvement of surface deformation monitoring precision of mining area.

Key words: Quantum neural network, Quadratic network fitting method, Deformation monitoring, BP neural network, Residual error, Internal precision and external precision