Research On Resource Scheduling Of Cloud Computing Based On Evolutionary Game

Cloud computing is a new business model for on-demand services that provides easy and fast access to all types of resources in the cloud.In cloud computing,task scheduling is the core issue.How to properly distribute the tasks submitted by users to the cloud computing data center is not only related to the user experience and service quality,but also has a vital impact on the service provision capabilities of cloud service providers,server cluster load balancing,and operational cost control..At present,Ant Colony Optimization(ACO)has a wide range of applications in cloud computing resource scheduling algorithms based on the advantages of pheromone volatilization mechanism.However,due to the positive feedback effect,the ant colony algorithm is easier to focus on the path where the local optimal solution is easy to find,but can not further explore the solution space of the problem.The manifestation is more obvious.This thesis focuses on the two aspects of optimal span and load balancing,and studies the problem of resource scheduling.The main work is as follows:(1)The improvement direction and ideas of the existing resource scheduling algorithms are studied.Aiming at the huge amount of data in the resource scheduling model under cloud environment,the resource-localization-first resource scheduling strategy based on evolutionary game is studied.In this thesis,the evolutionary game theory(EGT)is used to optimize the parameters of the ant colony algorithm and the update mechanism of the reward coefficient,so that it can be applied to the resource scheduling problem.The parameter optimization method based on Evolutionary Game Theory-Ant Colony Optimization(EGT-ACO)is used for task scheduling,which improves the resource locality of the task and reduces the data block transmission brought by the job scheduling process.Resource loss.This is more fair to each user in resource scheduling.(2)For the problem that the EGT-ACO algorithm does not consider the real-time performance of the virtual machine in the actual resource scheduling environment,the concept of reward coefficient is introduced to adjust the performance factors of the virtual machine processing capability and bandwidth in the resource scheduling model.influences.The conceptual modeling of reward coefficient is studied,and the method of updating reward coefficient according to three different strategies is proposed.Finally,three strategies are used as the operation of the evolutionary game model of the population.The ant colony algorithm based on hybrid strategy game(Mixed-ACO)is used to simulate the resource scheduling.After simulation,the experimental results show that the Mixed-ACO method can improve the matching degree between the task and the virtual machine and solve the load imbalance problem of the existing scheduling algorithm.Finally,the Ant Colony Algorithm resource scheduling model based on the evolutionary game method can improve the overall completion efficiency of the task in the scheduling process and reduce the waiting time of the task,and this effect is more obvious when the number of tasks increases.At the same time,tasks can be more fairly assigned to virtual machines during the scheduling process.