PBCOPSO: A Parallel Optimization Algorithm for Task Scheduling in Cloud Environment

Abstract

Objectives: Cloud environment requires scheduling of independent tasks with the available resources to minimize the total execution time and to optimize the resource utilization in cloud environment. Methods: Evolutionary algorithms are widely used to find the suboptimal solution of a problem. This work adopts a parallel approach that considers Bee Colony Optimization (BCO) in parallel with Particle Swarm Optimization (PSO) for cloud task scheduling. The proposed approach is named as Parallel Bee Colony Optimization Particle Swarm Optimization (PBCOPSO). Findings: The results show that the proposed approach minimizes Makespan with optimized resource utilization. It is observed that the proposed method improved resource utilization by an average of 5.0383% when compared with Min-Min algorithm and by an average of 3.7243% when compared with Improved Bee Colony Optimization (IBCO). Novelty of the Study: The proposed hybrid PBCOPSO enables improved search in the solution space due to the parallel execution of BCO and PSO leading to better final solution quality and lower execution time. Conclusion: Thus two metrics namely makespan and resource utilization are evaluated and an optimal task to resource mapping is achieved with hybridization.