A new modified firefly algorithm for function optimization

Abstract

This paper intends to provide a modified firefly algorithm based on firefly algorithm and improved particle swarm optimization. This firefly algorithm is a category of nature-enthused algorithm of swarm intelligence, i.e. depends on the response of a firefly to the light of other fireflies and also perform well on various numerical optimization problems. The modified algorithm uses the improved velocity concept of particle swarm optimization to enhance the searching behavior of standard algorithm. A comparison of the firefly algorithm with that of modified firefly algorithm is performed for some standard benchmark functions through simulations. The algorithms are also checked in various standard dimensions for providing effective output. The simulated results prove the superiority of modified firefly algorithm as compared to the traditional firefly algorithm in standard benchmark functions and in all dimensions. The results give an idea that the proposed modified algorithm enriches performance of the standard firefly algorithm and converges more quickly with less time to produce optimum solution.