ABSTRACT
ABSTRACT: Aiming at defect of premature convergence in particle swarm optimization algorithm, an improved cultural particle swarm optimization algorithm is put forward for load balancing of computing cluster. Firstly, main population space of particle swarm optimization algorithm and knowledge space of cultural algorithm are combined to form the mechanism of “double evolution and double promotion”, which can improve global searching capacity of algorithm and operational efficiency. Then evolution mechanism of genetic algorithm is adopted to enhance evolution operation of knowledge space. Finally, the algorithm is applied to load balancing problem of computing cluster to find solution. Experimental results show that the proposed algorithm increases resource utilization rate of computing cluster and makes load more balanced, proving to be an effective and reliable load balancing algorithm.
