Research on Virtual Machine Layout Strategy Based on Improved Particle Swarm Optimization Algorithm

Abstract

In the cloud computing center, the scheduling module allocates virtual machines to physical servers according to the virtual machine s resource usage, regardless of the physical server s overall and long-term resource utilization, which causes a large amount of energy loss in the cloud computing center. The virtual machine placement algorithm provides a way to save energy and improve resource management. This paper proposes a particle swarm optimization algorithm with crossover operator (CPSO) to maximize the use of resources and reduce energy consumption. In the article we designed a new fitness function, which optimizes the algorithm from three goals: load balancing, resource utilization and physical server usage. By adding the crossover operator in the genetic algorithm to the particle swarm optimization algorithm, the fitness value can be prevented from entering the local optimum too early. The algorithm can adaptively adjust the crossover probability and speed up the convergence of the algorithm. Finally, the algorithm is evaluated experimentally. The results show that CPSO is superior to the discrete particle swarm optimization (DPSO) and greedy algorithm (Best-Fit) in terms of resource utilization and physical machine usage. And the solution obtained by the algorithm is close to the optimal solution.