Abstract: A hybrid algorithm was proposed to solve function optimization problems in noisy environment which combined the Unified Particle Swarm Optimization Scheme, hypothesis test and optimal computing budget allocation technique together. The algorithm has good abilities of exploration and exploitation. Numerical simulations based on several representative benchmark problems were carried out in noisy environment and a comparison was made between UPSOOHT and several popular algorithms. Additionally, the influences of OCBA and noise magnitude were studied. The results show that UPSOOHT has a better performance.

Key words: unified particle swarm optimization, noisy environment, function optimization, hybrid optimization algorithm