Load Balancing of Virtual Machines in Cloud Computing Environment Using Improved Ant Colony Algorithm

Abstract

Load balancing of virtual machines is one of the most significant issues in cloud computing research. A common approach is to employ intelligent algorithms such as Ant Colony Optimization (ACO). However, there are two main issues with traditional ACO. First, ACO is very dependent on the initial conditions, which might affect the final optimal solution and the convergence speed. To solve this problem, we propose to employ Genetic Algorithm (GA) for ACO initialization. Second, ACO could arrive at local optimal point, and the convergence speed is typically low. Along this line, we introduce the idea of Simulated Annealing (SA) to avoid local optimal and accelerate the convergence. Lastly, our experiments show that our improved ACO achieves good performance in load balancing.