A modified Lévy flight distribution for solving high-dimensional numerical optimization problems

Lévy flight distribution is a recent meta-heuristic inspired by lévy flight random walk for exploring unknown large search spaces. Similar to other original metaheuristic algorithms, Lévy flight distribution can suffer from drawbacks, such as being trapped in minimum local areas and imbalance between the exploitation and exploration. To overcome these weaknesses and enhance the ability of Lévy flight distribution in solving high-dimensional numerical optimization problems, a modified Lévy flight distribution, called MLFD, is presented. Firstly, Lévy flight distribution has good exploration ability; secondly, the symbiosis organisms search has a strong exploitation capability in the mutualism phase. By introducing the mutualism phase, the exploitation ability of the algorithm is improved effectively and help avoid premature convergence. Moreover, a new differential variation strategy is proposed to enhance the diversity of the population and make the algorithm jump out of the local optimum in time. Seventeen well-known high-dimensional unconstrained problems are utilized to compare the proposed algorithm with other nine classical algorithms. The experimental results and statistical analysis demonstrate that MLFD algorithm has promising effectiveness and performance compared with other nine classical algorithms.