Mayfly Algorithm with Automatic Parameter Adaptation with Fuzzy Logic

Inspired by the behavior of ephemeral insects, the Mayfly algorithm was developed, considering their short lifespan and mating patterns for continuous evolution. It represents a significant improvement over the particle swarm optimization algorithm, combining the intelligence observed in these insects with evolutionary algorithms. A modification to its parameters using type-1 fuzzy logic was proposed to improve convergence efficiency. Mayfly exhibits good exploitation but poor exploration; however, hybridization with fuzzy parameter adaptation, using parameters a1 and mu, enhances its performance. This deviation from local optima improves the algorithm's diversity, favoring exploration. Out of the 10 chosen benchmark functions, Mayfly outperforms the original in 6 of them. Future work includes optimizing the membership functions of the fuzzy adapter using a genetic algorithm and incorporating an adapter with type-2 fuzzy logic. Such adaptations could further enhance the algorithm's performance and exploration capabilities.