Adaptive firefly algorithm based on reverse search strategy

Abstract

Firefly algorithm is proposed by Prof. Yang Xinshe for solving global optimization problems, which uses the principle of mutual attraction of fireflies in nature. The firefly algorithm is a branch of evolutionary algorithms, which is often used to solve single-objective global optimization problems with fewer parameters, easy to implement and easy to understand. However, the traditional firefly algorithm uses the full-attraction model in updating, which is easy to fall into local optimum. Therefore, a firefly algorithm that performs backward search with Poisson distributed probabilities is proposed, which enables the firefly to search more widely in the solution space and easily jump out of the local optimum. Comparative experiments are conducted on 28 functions of the CEC2013 test set. The experimental results show that in 22 of the functions the firefly with the reverse search strategy performs more accurately than the other improved firefly algorithms and in 26 of the functions the firefly with the reverse search strategy converges faster than the other fireflies.