Application Of MABC-SVR In Slope Reliability Analysis

The deformation and failure of slope is a key subject in the field of geotechnical engineering,the reliability analysis of slope engineering can provide a theoretical basis for engineering projects.Owing to the geological condition of the slope is highly nonlinear,it is difficult to describe the relationship between the safety factor and the influencing factors by a display function.Therefore,this paper combines the support vector machine with the Monte Carlo method,and then analyzes the reliability of the slope.The emphases are as follows:(1)In view of the disadvantage that the original artificial bee colony algorithm is easy to fall into the local optimum when searching,a modified artificial bee colony algorithm(MABC),is proposed in this paper.In this method,the fitness of honey source is sorted first,and finding the highest fitness source,then searching a better solution around it.The MABC has faster convergence speed and better overall performance than the original ABC algorithm.(2)Using the MABC to optimize support vector machine(MABC-SVR),and then establishing 80 slope safety factor prediction models based on this method,the prediction model has less error and higher accuracy,and the predicted value is highly fitted with the actual value.(3)Combining the Monte Carlo simulation method with the MABC-SVR model,analyzing the reliability of the typical slope and Qinghua Temple landslide before and after treatment,comparing reliability index and failure probability with other methods.the results are basically consistent.It is proved that the prediction model and reliability calculation method established in this paper are reasonable and feasible,and the accuracy is higher.