Augmented Brain Storm Optimization with Mutation Strategies

摘要

Brain storm optimization (BSO) is a recently proposed novel and promising swarm intelligence algorithm which models the human brainstorming problem-solving process. In BSO, the search areas are grouped into several clusters resulting in the diversity of population decrease in iterations. Hence, original BSO algorithm has suffered from low convergence speed and getting trapped into local optimum when solving global optimization problems since its inception. To address the issues, an augmented brain storm optimization with two mutation-based strategies (ABSO) is proposed in this study. First, a search technique based on non-uniform mutation is employed to accelerate the convergence speed of individuals locally. Second, a random mutation inspired by differential evolution is utilized to enhance the exploration capability globally. Finally, the performance of ABSO algorithm is tested on eighteen benchmark functions with various properties. Compared with the other algorithms, experimental results indicate that the proposed algorithm obviously enhance the performance of original BSO for global optimization in terms of solution accuracy and convergence speed.