A modified fireworks algorithm with dynamic search interval based on closed-loop control

Abstract

Fireworks algorithm performs better in solving some complex real-world engineering optimization problems, but it like other swarm intelligent algorithms, which also have the problems of slow search speed, has low efficiency and easy to fall into local optimum. Therefore, a modified fireworks algorithm is proposed for amending these weaknesses in this paper. Firstly, a closed-loop dynamic search interval adjustment strategy based on feedback control is proposed to improve search effective. Reference selection, controller design and center determination method are designed to obtain a reasonable dynamic search interval. Then we analyze the effective key parameters and design an explosion radius adjustment method related to iteration stage. Inspired by the scale law of foraging behavior of marine predators, Lévy flight and Brownian movement are applied to generate mutation sparks to enhance the algorithm local search capability. To make high-quality fireworks produce denser sparks, an explosion intensity operator based on the ranking of each firework is presented. Finally, a large number of experiments are used to verify the performance of the proposed algorithm, compared to other well-known algorithms, results confirm the superiority of this algorithm in terms of convergence rate and global search capability.