Research On Load Forecasting And Scheduling Strategy Of Cloud Computing Virtual Machine

The rapid development of modern society has brought the rise of the cloud computing industry.Cloud computing is becoming more and more popular as a new efficient and inexpensive computing model,and the problem of load resource prediction and resource scheduling strategy in cloud environment has been studied by scholars and has never been stopped.Cloud computing systems often result in waste of resources,unbalanced load,and low resource utilization,and these problems are getting more and more attention.This paper combines the load forecasting of virtual machine with the dynamic scheduling strategy of virtual machine,which can effectively improve the resource utilization ratio of cloud computing environment,solve the problem of resource waste and balance the resource utilization ratio,then bring some enlightenment to the deeper research.Firstly,studied the time-varying characteristics of virtual machine load,and the ARMA(p,q)model based on time series is used to predict the load value of virtual machine in real time.According to the load sequence data and R language,the forecasting model is ARMA(0,1,1),and the validity of the forecasting model is verified by model and parameter test of the residual.This model can be used to predict the load of the virtual machine in different time periods in the future,which provides a scientific basis for dynamic load adjustment in cloud data center.Secondly,in order to improve the resource utilization in cloud computing environment,a new resource utilization equilibrium model and a virtual machine task allocation minimum time model are proposed.Based on the basic particle swarm optimization algorithm,the inertia weight and learning factor are optimized,the optimized PSO algorithm is obtained and used to solve the model.Finally,two test functions defined by Matlab are used to simulate the two different algorithms,and the validity of the optimized algorithm is verified.Then the CloudSim cloud simulation platform is used to verify the feasibility of the algorithm and model from three aspects,and prove that the optimized PSO algorithm can find the optimal solution accurately and quickly.