Research On Virtual Machine Placement Method In Cloud Data Center Based On Improved Hybrid Optimization Algorithm

With the widespread use of cloud computing and virtualization technology,the deployment and management of virtual machines have become an important research direction.A virtual machine is a virtual computer running on a physical host that can improve server resource utilization and flexibility.Virtual machine placement is a key issue in virtualization environments,which involves how to map virtual machines to physical servers to achieve optimal resource utilization and performance.With the explosive growth of data center machines,a series of serious problems have emerged,such as high energy consumption and inefficient resource utilization.In this situation,how to place virtual machines reasonably to maximize the utilization of data center resources has become a very important research direction.Although there have been a large number of research results both domestically and internationally,we still need to further optimize virtual machine placement algorithms to meet the growing demand of data centers.Virtual machine placement can affect multiple aspects of virtualization environments,such as load balancing,resource utilization,and service quality,and thus has important practical application value.To address the issues of high energy consumption,load imbalance,and low resource utilization in cloud data centers,this article proposes two different virtual machine placement schemes.These schemes aim to optimize resource utilization efficiency in data centers,thereby cutting down energy consumption,and bring about height load balancing and resource utilization.The main research content of this thesis is grouped into two sections as following:(1)Addressing the problems of high energy consumption,load imbalance,and low resource utilization in cloud data centers,a virtual computer placement method based on hybrid NSGA-II particle swarm optimization algorithm is proposed.In the selection of optimization objectives,energy consumption and load balancing are selected as the optimization indicators for multi-objective optimization.Dynamic inertia weight is used to improve the search speed and exploration ability of particles in the search space.The proposed based on hybrid NSGA-II particle swarm optimization algorithm and several other algorithms are experimentally tested on the Cloud Sim simulation platform.The results show that the proposed NSGA-II integrated particle swarm hybrid algorithm has good convergence performance,effectively reduces energy consumption,balances loads,and improves resource utilization.(2)In order to further cut down the energy consumption of physical hosts in cloud data centers,balance the load among physical hosts,and improve resource utilization,a virtual computer placement method in view of the ACO fusion particle swarm optimization blending algorithm is proposed.Energy consumption and load balancing are still selected as optimization indicators,and an adaptive adjustment strategy is used to help the ACO fusion particle swarm algorithm dynamically adjust parameter values.Weighting the multi-objective optimization model into a single objective model,simplifying calculation and solving problems,and effectively handling constraints in virtual machine placement problems.Finally,experiments were conducted on the Cloud Sim simulation platform on the proposed ACO fusion particle swarm optimization blending algorithm and several other algorithms.The results indicate that the proposed ACO fusion particle swarm optimization blending algorithm can effectively cut down energy consumption and balance the load,and its complexity is not very high,making it easy to implement.Especially when facing a large number of virtual machine placement requests,the proposed algorithm performs even better.