An improved moth-flame algorithm based on differential evolution and shuffled frog leaping algorithm

Abstract

Generally speaking, in swarm intelligence algorithm, keeping the diversity of the population and strengthening its ability to escape the local optima will help to enhance the performance of an optimization algorithm. To strengthen these abilities, an improved moth-flame optimization (MFO) algorithm combining differential evolution (DE) and shuffled frog leaping algorithm (SFLA), namely DEFMFO, is proposed. In the proposed algorithm, the DE is used to enrich the population diversity, and the SFLA is proposed to get the optimal solution which strengthens the ability to escape the local optima. The simulation results show that, compared with particle swarm optimization (PSO), bat algorithm (BA) and classic MFO by testing on eight benchmark functions, the DEFMFO algorithm has better performance in global search and convergence speed.