The multi-start mayfly optimization algorithm

Abstract

Although there would be multiple ways for individuals in the mayfly optimization (MO) algorithm to update their velocities, the individuals would still be trapped easily in local optima. To reduce the probability being trapped, the mayflies might be reinitialized during iterations. In this paper, the multi-start methods was introduced to the MO algorithm and consequently, the male mayflies would be reinitialized same as the way at the beginning. Simulation experiments were carried out and results verifed that the multi-start MO algorithm would perform better in optimizing the benchmark functions than the original version.