Fuzzy-weighted differential evolution computing paradigm for fractional order nonlinear wiener systems

초록

The parameter estimation of fractional order nonlinear Wiener system is complex and challenging task due to the presence of significant nonlinearity at the output block, unknown fractional order, unknown parameters of the linear/nonlinear blocks, and the unmeasurable intermediate variables as well as the states. In this study, a novel design of fuzzy-weighted differential evolution is presented for parameter estimation of fractional order nonlinear Wiener (FO-NW) systems which are designed as an extension of conventional Wiener type models subjected to be fractional order system by exploiting Grunwald-Letnikov fractional derivative. The parameter estimation problem of FO-NW systems is constructed by defining a merit/error function between the true and estimated response via knacks of approximation theory in mean square error sense. Fuzzy Weighted Differential evolution algorithm is employed as an optimization mechanism to estimate the parameters of FO-NW systems with various output nonlinearities of polynomial, sinusoidal and sigmoidal kernels for low/high noisy environments in the system dynamics. Comparative studies based on rigorous statistics endorse the accurate, effective, stable and robust performance of fuzzy-weighted differential evolution algorithm. (c) 2022 Published by Elsevier Ltd.

키워드