A ranking-based firefly algorithm with adaptive parameter control and modified attraction

In this paper, an improved firefly algorithm (FA) is proposed for numerical optimisation. The new approach, termed RBFA, employs three strategies. First, each firefly is assigned a probability based on a ranking method. The probability determines whether a firefly moves to a new position. Then, a novel parameter method is designed to dynamically adjust the step size. Furthermore, a modified attraction strategy is used to strengthen the search efficiency. To verify the performance of RBFA, a set of famous benchmark problems are tested. Computational results demonstrate the proposed RBFA can obtain better performance than several other FA variants.