Research On Virtual Machine Initial Deployment Algorithm Based On Improved Genetic Algorithm

Cloud computing,from the distributed,pervasive computing,the use of virtualization,load balancing and other technologies to a higher flexibility,on-demand services and support personality configuration and other characteristics,to provide users with computing services and has achieved better Social and economic benefits.With the expansion of cloud computing,energy consumption is too large,resource utilization is not high,load imbalance and other issues become more prominent,how to improve the utilization of resources,balanced data center of the overall load,the need for further research.The initial deployment of the virtual machine is in the user request a certain service,the cloud computing data center to establish a task list,the task and virtual machine one by one,and then through the virtual machine resource management module,call a custom virtual machine deployment program,will have Virtual machines with multiple dimension resource requests are deployed to a physical machine with the appropriate resource capacity.Aiming at this kind of one-to-many mapping problem,the improved genetic algorithm is optimized to balance the data center load as the optimization target,and the fitness function based on load balancing is designed to select the standard of excellent individual,Search directions.In order to apply the genetic algorithm to the virtual machine deployment problem,the deployment scheme is abstracted and the reasonable real coding method is selected to facilitate the later operation.And the related genetic operators are improved to avoid the algorithm to complete the search prematurely,into the local optimal,improve the quality of the final solution.By using the crossover and mutation probability parameters of the evolution process,the dynamic adaptability of the algorithm is improved.By simulating the corresponding virtual machine initialization process on thesimulation platform CloudSim,the design and implementation of multiple contrast experiments are carried out.The experimental results show that the improved algorithm has excellent effect in balancing the data center load,and the feasibility is high.