Enhanced Moth-flame optimizer with mutation strategy for global optimization

Abstract

Moth-flame optimization (MFO) is a widely used nature-inspired algorithm characterized by a simple structure with simple parameters. However, for some complex optimization tasks, especially the high dimensional and multimodal problems, MFO may have problems with convergence or tend to fall into local optima. To overcome these limitations, here a series of new variants of MFO are proposed by combining MFO with Gaussian mutation (GM), Cauchy mutation (CM), Lévy mutation (LM) or the combination of GM, CM and LM. Specifically, GM is introduced into the basic MFO to improve neighborhood-informed capability. Then, CM with a large mutation step is adopted to enhance global exploration ability. Finally, LM is embedded to increase the randomness of search agents’ movement. The best variant of MFO was compared to 15 state-of-the-art algorithms and 4 well-known advanced optimization approaches on a comprehensive set of 23 benchmark problems and 30 CEC2017 benchmark tasks. The experimental results demonstrate that the three strategies can significantly boost exploration and exploitation capabilities of the basic MFO.