Empowering particle swarm optimization algorithm using multi agents' capability: A holonic approach

A novel particle swarm optimization algorithm based on holonic structure in multi agent systems is presented. The proposed algorithm employs holonic structure of multi agent systems for optimizing numerical functions. Paying more attention, particle swarm optimization algorithm and multi agent systems are similar in first glance. Their similarities are based on the fact that both of them are population based and do their tasks cooperatively. In proposed approach, PSO is considered as a multi agent system and particles as agents. Multi agent systems use organizational design because of quantitative effect on their performance. One of these organizations is holonic structure. By using this structure, particles are arranged in different groups or holons and make a holarchy. In this holarchy, different groups or holons can communicate with each other in order to search space more efficiently, avoiding premature convergence and trapping in local optimums. Proposed structure helps PSO to maintain particles' diversity and also makes a suitable balance between exploration and exploitation. The proposed algorithm is tested on a set of well-known test functions. Results have shown that the proposed algorithm is efficient, more accurate and outperforms other particle swarm optimization algorithms examined in this paper.