A simple diversity guided firefly algorithm

The purpose of this paper is to present a modified firefly algorithm (FA) considering the population diversity to avoid local optimum and improve the algorithm’s precision.

When the population diversity is below the given threshold value, the fireflies’ positions update according to the modified equation which can dynamically adjust the fireflies’ exploring and exploiting ability.

A novel metaheuristic algorithm called FA has emerged. It is inspired by the flashing behavior of fireflies. In basic FA, randomly generated solutions will be considered as fireflies, and brightness is associated with the objective function to be optimized. However, during the optimization process, the fireflies become more and more similar and gather into the neighborhood of the best firefly in the population, which may make the algorithm prematurely converged around the local solution.

Due to different dimensions and different ranges, the population diversity is different undoubtedly. And how to determine the diversity threshold value is still required to be further researched.

This paper presents a modified FA which uses a diversity threshold value to guide the algorithm to alternate between exploring and exploiting behavior. Experiments on 17 benchmark functions show that the proposed algorithm can improve the performance of the basic FA.