Many-Objective Brain Storm Optimization Algorithm

In recent years, many evolutionary algorithms and population-based algorithms have been developed for solving many-objective optimization problems. Inspired by the human brainstorming conference, Brain Storming Optimization (BSO) algorithm was guided by the cluster centers and other individuals with probability, which can balance convergence and diversity greatly. In this paper, the authors propose a novel brain storm optimization algorithm for many-objective optimization problem. The algorithm adopts the decision variable clustering method to divides the variables into convergence-related variables and diversity-related variables. The decomposition strategy is designed to increases selection pressure for the convergence-related variables, while the reference points strategy is adopted for the diversity-related variables to update the population and increase the diversity. Experimental results show that the proposed many-objective brain storm optimization algorithm is a very promising algorithm for solving many-objective optimization problems.