Optimal Choice of Parameters for Fireworks Algorithm

Abstract The Fireworks Algorithm (FWA) is a relatively novel metaheuristic algorithm based on simulating the flare-up process of fireworks exploding and illuminating the night sky. The effects of major parameters on FWA were investigated on the unimodal and multimodal benchmark functions. The parameters such as number of fireworks, explosion amplitude, number of explosion sparks and number of iterations, have significant impact on the performance of FWA. The inappropriate selection of these parameters may lead to wrong results. The characteristics of FWA's parameters are depicted and general recommendation for deciding the optimal parameter values is given in this paper.