Acceleration for Fireworks Algorithm Based on Amplitude Reduction Strategy and Local Optima-Based Selection Strategy

摘要

We propose two strategies for improving the performance of the Fireworks Algorithm (FWA). The first strategy is to decrease the amplitude of each firework according to the generation, where each firework has the same initial amplitude and decreases in size every generation rather than by dynamic allocation based on its fitness. The second strategy is a local optima-based selection of a firework in the next generation rather than the distance-based selection of the original FWA. We design a set of controlled experiments to evaluate these proposed strategies and run them with 20 benchmark functions in three different dimensions of 2-D, 10-D and 30-D. The experimental results demonstrate that both of the two proposed strategies can significantly improve the performance of the original FWA. The performance of the combination of the two proposed strategies can further improve that of each strategy in almost all cases.