From Resampling to Non-resampling: A Fireworks Algorithm-Based Framework for Solving Noisy Optimization Problems

摘要

Many resampling methods and non-resampling ones have been proposed to deal with noisy optimization problems. The former provides accurate fitness but demands more computational resources while the latter increases the diversity but may mislead the swarm. This paper proposes a fireworks algorithm (FWA) based framework to solve noisy optimization problems. It can gradually change its strategy from resampling to non-resampling during the evolutionary process. Experiments on CEC2015 benchmark functions with noises show that the algorithms based on the proposed framework outperform their original versions as well as their resampling versions.