Research On Particle Swarm Optimization Algorithm For Cloud Resource Scheduling In Uncertain Environment

Cloud resource scheduling is a key issue in the field of cloud computing.How to obtain an effective cloud resource scheduling scheme is a difficult problem that needs to be solved.Many scholars at home and abroad have established a variety of cloud resource scheduling models and used a variety of intelligent optimization algorithms to solve them.However,considering the existence of uncertainty in the actual process,in order to make the established cloud resource scheduling model closer to reality,based on the fuzzy programming theory,a fuzzy cloud resource scheduling model under uncertain environments is established,and based on the existing algorithms,a variety of optimized scheduling algorithms with better performance are proposed.Using triangular fuzzy numbers to represent the execution time of tasks on virtual machine resources,a fuzzy multi-objective cloud resource scheduling model under time-cost constraints is established.The mapping relationship between tasks and virtual machine resources is used as a decision variable.The goal of scheduling is in order to make the total execution time and total execution cost of the task reach the Pareto optimal solution.This paper proposes an improved particle swarm optimization algorithm(Re-randomization inertia weight orthogonal initialization particle swarm optimization,RIOPSO)to solve the model.The algorithm uses orthogonal initialization to improve the quality of the initial solution,using re-randomization and the strategy of updating the inertia weight in real time enables the particles to fully search in the solutio n space and finally converge to the optimal solution.The simulation experiment shows that it is necessary to consider the uncertainty.In the comparison experiment of the algorithm,it verifies the superior performance of the RIOPSO algorithm in the optimization and convergence speed when solving the cloud resource scheduling problem.Aiming at the uncertainty of cloud resource scheduling,triangular fuzzy numbers are used to fuzz the execution time and execution cost of tasks,and a fuzzy multi-objective scheduling model based on time-cost-load balance is established.A hybrid particle swarm optimization algorithm(CMA-PSO)is proposed to solve the model.The algorithm introduces the covariance matrix adaptive evolution strategy,so that the CMA-PSO algorithm has an initial value closer to the optimal solution at the initial stage,and evenly distributed around the optimal solution to improve the accuracy of the initial solution.Through simulation experiments,the optimization ability of CMA-PSO algorithm is compared and analyzed on multiple performance indicators using advanced algorithms and CMA-PSO algorithm.The experimental results prove the optimization of CMA-PSO algorithm on multiple performance indicators.Ability and convergence ability.In order to test and compare a variety of cloud resource scheduling models and algorithms,a cloud resource scheduling simulation software was established using the Swing graphical interface development kit.This software is versatile and extensible,and can integrate deterministic and uncertain environments.Cloud resource scheduling model,and can integrate PSO algorithm,RIOPSO algorithm,CMA-PSO algorithm and other intelligent optimization algorithms.Through convenient interface operation,intuitive simulation experiments and results can be displayed on different cloud resource scheduling models and algorithms,and the experimental results can be retained to provide experimental basis for the performance analysis of models and algorithms.