Point estimation of nonlinear regression model parameters with Genetic-Simplex hybrid algorithm

In this study, a Genetic-Simplex hybrid algorithm, which is composed of advantageous aspects of derivative-free optimization algorithms, such as Nelder-Mead Simplex (NMS) algorithm and Genetic Algorithm (GA), is used to obtain point estimates of nonlinear regression model parameters. In addition, it is studied to decide optimal values of GA tuning parameters by using Taguchi experimental design. In the study, point estimates of the parameters of a negative-exponential regression model, defined in the literature in accordance with a data set, are obtained by using the proposed optimization approaches. When the obtained results are compared with the results given in the literature, it is seen from the comperative results that estimates of model parameters are easily obtained and consistently approximated to the minimum value of the objective function by using Genetic-Simplex hybrid algorithm