Utilizing Multi-Agent Systems Approach in Firefly Algorithm

Generally, using collective intelligence is one of the interesting topics is researchers of recent years, which its purpose is modeling creatures' simple behaviors and their interaction with the environment and neighbor creatures to obtain more complex behaviors. We could utilize algorithms based on collective intelligence to solve complicated problems like optimization problems. So far, various algorithms have been purposed in this field which firefly algorithm is a variant of these. In this algorithm, each member acts as a better response concerning itself. However, this algorithm has some drawbacks like the consistency of parameters value, lack of balance between local search and global search, and others. On one hand, multi-agent systems are software systems that contain sets of agents. These agents perform their tasks together to solve a problem and reach the desired purpose. In this paper, we have tried to utilize a multi-agent system, in addition to meta-heuristic optimization algorithms, to improve the performance of the firefly algorithm to better cooperate warms populations with each other. The results of the experiment show the acceptable performance of the proposed algorithm.