Research On SVM Slope Deformation Prediction Optimized By Improved Quantum Particle Swarm Optimization

In recent years,while people continue to pursue economic development,there are more and more slope projects.In order to ensure people's lives and property safety,it is necessary to make reasonable early warning and forecast of possible landslide hazards caused by slope deformation.Support vector machine has obvious advantages in solving nonlinear problems such as slope deformation and displacement.Aiming at the parameter selection problem of traditional support vector machine model,an improved support vector machine model of quantum particle swarm optimization is proposed(IQPSO)and applied In the example of slope deformation prediction.In this paper,first of all,for the selection of kernel functions in the support vector machine model,a domestic actual slope case is used as the research object,different kernel functions are selected through experiments to perform fitting experiments,and the radial basis kernel is used in the prediction of slope deformation The function works better.Secondly,in view of the premature phenomenon of the parameter optimization process of the quantum particle swarm optimization algorithm,two improvements are made in this paper.First,because the selection of the contraction-expansion coefficient has a great influence on the accuracy of parameter optimization,in the optimization process,the value in the early stage needs to be larger and the rate of decrease is slow,while in the middle and later stages,the speed of the optimization value needs to be reduced.It is faster,so the contraction-expansion coefficient is changed by the nonlinear decreasing method instead of the traditional linear decreasing method of the quantum particle swarm;second,after the average optimal position is obtained,the evolution factor is added to make the algorithm jump out of the local optimal in the later stage.Continue Search for a better solution.Through simulation experiments,comparing the optimization results and accuracy,it is concluded that the IQPSO algorithm is superior to the QPSO and PSO algorithms in all aspects.Then,according to the optimization of IQPSO algorithm,the parameters and penalty parameters in the kernel function are obtained,and the SVM slope deformation prediction model based on IQPSO optimization is constructed.Finally,the constructed model was applied to the Dafengwan slope and Fujiapingzi slope of Xiluodu Hydropower Station,and compared with the prediction results of QPSO-SVM model and PSO-SVM model,using mean square error(MSE)and average relative error(MRE)evaluate the prediction results.Experimental results:(1)For the slope of Dafeng Bay,the MSE of PSO-SVM model,QPSO-SVM model and IQPSO-SVM model are 1.01,0.77 and 0.10,respectively,while the MRE of the three models are 3.61%,3.19% and 1.03%,(2)For Fujiapingzi slope,the MSE of PSO-SVM model,QPSO-SVM model and IQPSO-SVM model are 4.20,0.92 and 0.15 respectively,while the MRE of the three models are 1.65% and 0.77 respectively % And 0.30%.From the experimental results,it can be concluded that the IQPSO-SVM model has improved prediction accuracy and better results than the QPSO-SVM model and the PSO-SVM model in slope deformation prediction.