Research on the selection method of multi-VM resource adjustment strategy in a single PM based on genetic algorithm

The selection method of resource adjustment strategy is a key step of multi-VM (Virtual Machine) resource adjustment in a single physical machine (PM). The traditional genetic algorithm (GA) do not evaluate and filter the initial population, and not make full use of decision of historical data as to increase the optimal solution time either. Based on the conventional research, this paper establishes the relation model between the service performance and the amount of resources consumption (P-R model), which is used to evaluate and filter the initial population, and presents design method of the revenue function and the termination conditions. It also presents the way which turns the empirical data into historical decision and uses it in the next cycle. The experiment results indicate the method is able to maintain high resource utilization and meets the demands of response time.