A Cooperative Multi-swarm Particle Swarm Optimizer Based Hidden Markov Model

Particle swarm optimization (PSO) is a population-based stochastic metaheuristic algorithm; it has been successful in dealing with a multitude of optimization problems. Many PSO variants have been created to boost its optimization capabilities, in particular, to cope with more complex problems. In this paper, we provide a new approach of multi-population particle swarm optimization with a cooperation strategy. The proposed algorithm splits the PSO population into four sub swarms and attributes a main role to each one. A machine learning technique is designed as an individual level to allow each particle to determine its suitable swarm membership at each iteration. In a collective level, cooperative rules are designed between swarms to ensure more diversity and realize the better solution using a Master/Slave cooperation scheme. Several simulations are performed on a set of benchmark functions to examine the performances of this approach compared to a multitude of state of the art of PSO variants. Experiments reveal a good computational efficiency of the presented method with distinguishable performances. © 2020, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG.