Using Population Migration and Mutation to Improve Loser-Out Tournament-Based Fireworks Algorithm

The fireworks algorithm (FWA) is a newly proposed swarm intelligence algorithm inspired by the phenomena of fireworks explosion and has solved many real-world optimization problems successfully. A loser-out tournament-based fireworks algorithm (LoTFWA) is a new baseline in the development of FWA due to its outstanding independent framework and competition mechanism for multimodal optimization. Under this framework, each firework calculates its expected fitness improvement compared with the best fitness to determine whether to be reinitialized. Although LoTFWA achieves the best performance among the variants of FWA, it lacks of comprehensive consideration of the fireworks’ cooperation and hence weakens the algorithm’s power. This paper improves the cooperation of fireworks in LoTFWA based on the idea of population migration and mutation in biogeography-based optimization (BBO). The proposed mechanism not only promotes fireworks’ exploration ability but also enhances their exploitation ability greatly. Experimental results show that the proposed algorithm attains superior performance than the state-of-the-art fireworks algorithm in both unimodal and multimodal functions.