Parameter estimation of Richards model based on improved drosophila algorithm

Abstract

Aiming at the practical problem of complex parameter estimation in Richards model, the parameter estimation problem of Richards model is transformed into a multi-dimensional nonlinear function optimization problem by following the principle of least squares. Combined with the deformation monitoring data of foundation pit slope in a certain construction project and through C++ programming, variable random search step size strategy and variable drosophila population size strategy were introduced to realize the adaptive improvement of standard drosophila algorithm (FOA), so as to improve the local search ability and search efficiency of the algorithm. In order to verify the accuracy and effectiveness of the algorithm's estimation parameters, the improved Fruit fly algorithm (IFOA), particle swarm optimization (PSO) and genetic Algorithm (GA) are compared and analyzed. The results show that the improved Fruit fly algorithm (IFOA) is more applicable to the parameter estimation of Richards model.