Prediction of Soft Soil Roadbed Settlement Based on GA-SVM Model
In order to realize the settlement prediction of highway subgrade under soft soil geological conditions,a subgrade settlement prediction model based on Genetic Algorithm optimized Support Vector Machine is proposed.The basic model of soft soil subgrade set-tlement prediction is constructed by using Support Vector Machine algorithm,and the super parameters of the model are optimized by u-sing Genetic Algorithm.Taking the monitoring data of highway settlement observation points as samples,the subgrade settlement predic-tion model based on GA-SVM is established.The results show that Genetic Algorithm can effectively improve the fitting accuracy of Support Vector Machine for settlement data;The average error and root mean square error of GA-SVM settlement prediction model for 10 validation samples were 0.001 5 mm and 0.015 5 mm,respectively;The subgrade structure at the observation points in the next 10 periods tends to be stable,and the average predicted settlement after stabilization is about 0.03 mm/d.
万方数据
维普
