Adaptive Fireworks Algorithm Based on Simulated Annealing

Abstract

This paper proposes an adaptive fireworks algorithm based on simulated annealing to solve global optimization problems. It normalizes the individuals before the explosion operation and then restores them after the explosion operation. The subpopulations are produced simultaneously by the explosion operation and the mutation operation, and then a new way to produce the explosion amplitude by introducing the simulated annealing is used to adjust the explosion amplitude adaptively along with the evolution. Experimental results on eight benchmark functions with different shift values show that the proposed algorithm has higher accuracy and faster convergence rate than the other five compared algorithms.