AZERBAIJAN NATIONAL ACADEMY OF SCIENCES
az en ru
MULTICRITERIA OPTIMIZATION METHOD FOR LOAD BALANCING IN CLOUD COMPUTING
Rasim M. Alguliyev, Yadigar N. İmamverdiyev, Fargana C. Abdullayeva

Optimizing of the task scheduling process in the cloud environment is a multicriteria NP-hard problem. In this paper, weighted load balancing method (αPSO – TBLB) based on PSO algorithm is proposed. The method provides optimal migration of tasks from the loaded virtual machines to the less loaded virtual machine to prevent the excessive load in virtual machines of the cloud infrastructure. In the proposed optimization method, the minimization of the processing time of tasks and the transfer time of tasks were selected as the target functions. Experimental testing of the proposed approach was carried out in the Jswarm and Cloudsim programs. As a result of the simulation on the basis of the proposed method, an optimal solution for task scheduling was found, uniform distribution of tasks in virtual machines (VMs) was provided. Moreover, in the process of assigning tasks to virtual machines, a minimal time consumption was achieved (pp.3-13). 

Key words: cloud computing, Particle Swarm Optimization (PSO), virtual machine migration, task sheduling, Cloudsim, Jswarm, data intensive, computing intensive.