Penaksiran Parameter Regresi Nonlinear Menggunakan Particle Swarm Optimization dan Genetic Algorithm

Abstract

This research aims to estimate and compare the results of parameter estimation for nonlinear regression using the Particle Swarm Optimization and Genetic Algorithm algorithms. In Nonlinear Regression Models, parameter estimation is carried out using the conventional nonlinear strategy referred to as "Nonlinear Least Squares Regression." When estimating nonlinear regression model parameters, the Gauss-Newton Method and the Levenberg-Marquadrt Method still do not guarantee convergence and a global optimum, despite the fact that PSO and GA have provided guarantees for one. The distinction in the wellness esteem coming about because of these two methodologies is utilized to evaluate it. After simulating the estimator's data with the Matlab program, it was discovered that the genetic algorithm was more effective for parameter estimation in nonlinear regression models. The results show that the standard nonlinear least squares regression fits GA and PSO well. GA parameter estimation, on the other hand, shows greater success.