Peak Searching for Multimodal Optimization through Gradient-based Stochastic Extremum Seeking Technique

Abstract

In this paper, a global peak searching algorithm is proposed for multimodal optimization problem. The proposed algorithm is built through combining gradient-based stochastic extremum seeking method with iterative approach. By increasing the initial search point, the gradient-based stochastic extremum seeking algorithm with different initial condition is iteratively operated to seek a list of the local extreme values near different initial points. Then, the global peak can be found by comparing the list of the local extreme values. The stability anlysis and working process of the global peak searching algorithm is proven. Simulation results are given to show the effectiveness of the proposed control method.