Research On Virtual Machine Adaptive Migration Strategy Based On Reinforcement Learning

The management of data centers is a hotspot in the cloud computing industry.How to improve the resource utilization rate of the data center and ensure the load balance of the data center has become the top priority of the research.At present,the optimization of cloud data center resource utilization rate is mainly accomplished through virtual machine migration technology.However,the existing virtual machine migration strategy is difficult to take into account the two goals of energy saving and data center load balancing.Therefore,considering the power consumption and load balancing of the data center,establishing an effective virtual machine migration model is an effective way.By studying the existing data center management strategies,the research results obtained in this thesis mainly include:1.Aiming at the problem that the existing single prediction model is difficult to accurately predict,the Induced Ordered Weighted Average(IOWA)operator is introduced.The operator weights the accuracy of the fitting at each time point in the sample interval by each single prediction method in order,and establishes a new load combination prediction model based on the sum of squared errors.After a large number of numerical examples verify the correctness and effectiveness of the model.2.In view of the current virtual machine migration model does not comprehensively consider the problem of energy-saving data center load balancing,this thesis establishes a data center reinforcement learning model.Through the method of reinforcement learning,explore the best strategy for virtual machine migration in a cloud environment,reduce the number of virtual machine migrations,and achieve load balancing in a low-power data center.3.This thesis use a visual method to design and implement a data center load balancing system.The two models proposed in this thesis are verified,and a new solution is proposed for the data center load balancing strategy.