Quality of service-based particle swarm optimization scheduling in cloud computing

Abstract

Scheduling is important in cloud computing system. In this paper, an adaptive particle swarm optimization (PSO) algorithm is proposed to optimize quality of service (Qos)-guided task scheduling in cloud computing. This scheduling targets a trade-off between completion time and cost. The proposed algorithm adaptively changes PSO parameters according to the evolution state evaluation. This adaptation can avoid premature convergence and explore the search space more efficiently. When swarms are trapped into premature convergence, mutation is introduced to the velocity and position updating strategy to improve the ability of global search. Simulation results reveal that the algorithm can achieve significant optimization of completion time and cost.