Application Research Of Group Teachinglearning Based Multi-objective Optimization Algorithm For Virtual Machine Scheduling In Cloud Computing

By utilizing virtualization technology,cloud computing creates a virtual pool of physical resources that may be homogeneous or heterogeneous.Cloud users can purchase corresponding services based on their own requirements.Nevertheless,as cloud computing technology advances and more individuals utilize cloud services,the size of cloud data centers is growing.One of the pressing issues in cloud computing is the efficient scheduling of virtual machines in a vast and intricate cloud infrastructure,aiming to enhance the scheduling performance of the cloud platform and cater to user requirements.With the increasing number of tasks,the virtual machine scheduling problem in cloud platforms becomes highly intricate and is considered an NP-hard problem.The performance of cloud platforms heavily relies on the design of the virtual machine scheduling algorithm.To enhance load balancing and reduce resource wastage,this study delves into the virtual machine scheduling algorithm and its practical application.The primary focus and novelty of this thesis are outlined below:(1)A teaching-learning based optimization algorithm has been suggested,which involves grouping individuals based on their fitness levels.The current teaching-learning based optimization algorithms have certain drawbacks,including sluggish convergence rates and insufficient search capabilities.This study employs the pedagogical principle of "teaching students according to their aptitude" to group the populations in the algorithm.Each group adopts different strategies for iterative optimization in the teacher stage and the student stage respectively,so as to improve the overall optimization ability of the population.In the teaching stage,teachers no longer teach according to class average grades,but according to each student’s grades individually;In the student stage,students no longer choose study partners randomly in the whole class,but choose study partners in limited groups.The experimental results on CEC test set show that compared with some existing evolutionary algorithms,the grouping teaching-learning based optimization algorithm has better optimization ability.(2)Propose a group teaching-learning based multi-objective optimization algorithm for virtual machine scheduling.The traditional virtual machine scheduling strategies suffer from issues such as physical host resource wastage and load imbalance.To address these problems,we formulate a multi-objective virtual machine scheduling optimization model and design an algorithm based on group teaching-learning iterative optimization strategy.This algorithm enhances the search capability of the optimal mapping scheme between virtual machines and physical hosts for improved performance.Based on cloud computing simulation environment CloudSim,the virtual machine scheduling comparison test was carried out.Experimental results show that compared with the existing evolutionary algorithm,TLBO algorithm,improved multi-operator differential evolution algorithm and other algorithms,the proposed group teaching-learning multiobjective optimization algorithm has a more balanced load on the physical host,which verifies the effectiveness and performance of the algorithm.(3)A cloud computing virtual machine management system was developed and implemented,utilizing multi-objective optimization of group teaching-learning scheduling strategy.This system is built upon the open source cloud computing project OpenStack.Through functional testing,it has been demonstrated that the virtual machine management system can effectively manage and schedule virtual machines on the cloud platform in a unified manner.Additionally,the system boasts a user-friendly interface and easy operation.It can meet the requirements of high-performance scheduling management on the cloud platform,and has strong practicability.