A hybrid fireworks optimization method with differential evolution operators

Abstract

Fireworks algorithm (FA) is a relatively new swarm-based metaheuristic for global optimization. The algorithm is inspired by the phenomenon of fireworks display and has a promising performance on a number of benchmark functions. However, in the sense of swarm intelligence, the individuals including fireworks and sparks are not well-informed by the whole swarm. In this paper we develop an improved version of the FA by combining with differential evolution (DE) operators: mutation, crossover, and selection. At each iteration of the algorithm, most of the newly generated solutions are updated under the guidance of two different vectors that are randomly selected from highly ranked solutions, which increases the information sharing among the individual solutions to a great extent. Experimental results show that the DE operators can improve diversity and avoid prematurity effectively, and the hybrid method outperforms both the FA and the DE on the selected benchmark functions.