Um Algoritmo Híbrido Baseado em Fireworks e Evolução Diferencial para Otimização Numérica

This paper presents a cooperative hybrid algorithm based-on two metaheuristics: Fireworks and Differential Evolution. The main idea is to randomly choose which metaheuristic will be used on each iteration during the optimization process. We adopted the Rosenbrock function in our preliminary experiments. It is a challenging function to optimize because it is multimodal and non-separable when solving it in more than three dimensions. The initial investigations show that the hybrid algorithm tends to present better results than the canonical ones.