改进布谷鸟算法优化支持向量机的隧道变形预测分析

Research on tunnel deformation prediction by support vector machine model based on improved cuckoo search algorithm

传统支持向量机参数寻优时间较长,模型的稳定性也较差。为解决传统支持向量机预测模型在隧道围岩变形预测中精度较低的问题,引入布谷鸟算法,同时改进布谷鸟算法中的搜索公式,从而提出一种基于改进布谷鸟算法的支持向量机组合预测模型。将此模型运用到云南省昆明市阳宗隧道围岩位移预测中,并与标准布谷鸟算法的支持向量机组合模型以及单一的支持向量机模型比较,实验结果表明:基于改进布谷鸟算法的支持向量机模型精度更高,更具有优势。

The traditional support vector machine needs a long time to optimize parameters and the stability of the model is poor.In order to solve the problem of low accuracy of traditional support vector machine in tunnel surrounding rock deformation prediction,therefore,this paper combines support vector machine with cuckoo algorithm,and improves the search formula of cuckoo algorithm at the same time,by proposing a combined prediction model of support vector machine based on improved cuckoo algorithm.Finally,the model is applied to the displacement prediction of a tunnel surrounding rock.Comparing with the support vector machine combination model based on the standard cuckoo algorithm and the single support vector machine model,the experimental result shows that the support vector machine model based on the improved cuckoo algorithm has higher precision and more advantages.