Study on improved cuckoo search algorithm applied in parameters estimation of multi‐layer soil: 1D and 3D

Abstract

In this study, an improved cuckoo search (ICS) algorithm is presented to solve the parameter estimation problem of soil structure. For 1D soil parameters estimation, the objective function is constructed based on the relative root-mean-square error between measured and calculated soil apparent resistivity curves. The ICS algorithm extends the search range of soil model parameters through variable control parameters. Optimal resistivity and thickness of each layer are obtained by several different algorithms under the same test conditions based on two known four-layer soil models without noise. The results show that the ICS algorithm performs better in minimising error. This study also tests the optimisation performance of the ICS algorithm in two actual high-quality data cases, comparative results of the ICS algorithm and existing literature validate the effectiveness and accuracy of ICS. Moreover, in order to make soil estimation results more visible in a specific area, the combination of the proposed estimated algorithm and a new arrangement way of measurement lines is proposed to estimate 3D soil electric parameter structure in Chengdu, China.