PERFORMANCE ENHANCEMENT OF THE DIFFERENTIAL EVOLUTION ALGORITHM USING LOCAL SEARCH AND A SELF-ADAPTIVE SCALING FACTOR

摘要

This paper presents a novel differential evolution (DE) algorithm using a dynamic strategy, local search, and a self-adaptive scaling factor (DELSbp) to enhance the performance of the traditional DE algorithm. The DELSbp consists of a dynamic strategy, which updates the optimal vector instantaneously, and back-propagation-based local search, which enhances its searching capability from the corresponding neighborhood. In addition, a self-adaptive scaling factor strategy, developed using a fuzzy logic system, is introduced to accelerate the convergence velocity. The inputs of fuzzy systems incorporate the change in fitness values and generation number to calculate a change in the scaling factor. The performance of the DELSbp algorithm has been demonstrated using experimental results from a set of standard test functions and through the estimation of coefficients for an IIR filter with noise. These results demonstrate the performance and efficacy of the DELSbp algorithm.