Partial population restart of firefly algorithm using complex network analysis

Abstract

In this paper, we are presenting a method for controlling population diversity of popular metaheuristic algorithm, which is the Firefly Algorithm (FA). Presented method is using the advantages of complex networks and their several characteristics, that can be helpful for the detailed analysis of metaheuristic algorithm inner dynamic. However, this application can be quite tricky, since there are too many possibilities, as to how to use the advantages of complex networks analysis approach. Through this work, we are trying to present a simple workflow for building and analysis of network and the most suitable choices in each step to achieve better results of FA, especially, when focusing on population diversity.