Cloud Task Scheduling Based on Improved Particle Swarm optimization Algorithm

Abstract

Aiming at the problem of task scheduling in cloud computing resource scheduling, a scheduling strategy combining genetic algorithm (GA) and improved particle swarm optimization algorithm (GA-IPSO) is proposed. Firstly, a multi-objective evaluated model is established considering the task completion time, maximum completion time and load balance. Secondly, GA is used to optimize the randomly generated solution space to generate the basic solution. Finally, the improved particle swarm optimization algorithm is proposed to obtain the optimal solution of cloud task scheduling. In this paper, particle swarm optimization (PSO) is improved by establishing nonlinear negative correlation between inertia weight and iteration times and combining individual cognitive learning factors with evaluation function values. Simulation results show that GA-IPSO reduces the fitness value, maximum completion time, task completion time and load balancing degree of virtual machines by 12.8%, 15.3%, 12.0%, 50.8% on average in small-scale tasks and by 18.9 %, 25.3 %, 15.6 %, 41.8 % on average for large-scale tasks compared with other algorithms.