An Improved Firefly Algorithm with Adaptive Strategies

摘要

Firefly Algorithm (FA) is a powerful swarm intelligence algorithm inspired by the flash phenomenon of the fireflies. However, it has weaknesses on optimizing high-dimensional problems. This paper presents an improved FA named Adaptive Firefly Algorithm (AFA). In the new algorithm, three strategies are proposed to improve its adaptability and overcome its weaknesses. The algorithm is tested on a set of benchmark functions with different dimensions. Canonical FA and two other well-known heuristic algorithms are also employed for comparison. Experiment and statistical results show that the AFA algorithm offers significant better performance than original FA algorithm. Especially on high-dimensional problems, it is superior to all other compared algorithms.