Research On Cloud Task Scheduling Strategy Based On Improved Simulated Annealing Algorithm

With the rapid development of the Internet,the number of users connecting to the cloud is increasing.Cloud computing providers use virtualization technology to virtualize a physical host into multiple virtual machines.Only a limited number of physical hosts can be used to ensure multiple users.A large number of servers are used to form a large-scale data center to meet the user’s task requirements.How to efficiently allocate resource nodes to users to meet their task requirements in cloud computing is particularly important.Simulated Annealing(SA)is a common algorithm for solving scheduling problems,but it has the problem of low efficiency,and it is difficult to obtain a high-quality final deployment solution.In order to solve this problem,this dissertation introduces the concept of immune mechanism into simulated annealing algorithm,and proposes a simulated annealing algorithm based on immune mechanism.In addition,the advantages of Tabu Search algorithm(TS)is combined with the advantages of improved simulated annealing algorithm.The above two improved simulated annealing algorithms are used in cloud computing resource scheduling scenarios.The research work and content of the dissertation are as follows:(1)This dissertation studies the cloud platform resource scheduling model,analyzes the hierarchical categories of resource scheduling in cloud computing,understands the characteristics of resource scheduling in cloud computing,mathematically models the resource scheduling problem in the cloud environment,and simulates resource scheduling problems as tasks mapping to virtual machine problems.Considering the usage of system resources and the running time of virtual machines,a multi-objective evaluation model including system completion time,system load balancing and system execution cost is constructed based on the traditional model.Aiming at this model,a simulated annealing algorithm based on immune mechanism is proposed to solve it.Through the immune mechanism,the resource scheduling target and the final allocation plan are corresponding to antigens and antibodies,respectively,and the Metropolis criterion is introduced in the process of antibody group update to accept inferior antibodies with a certain probability,which increases the diversity of the population and prevents the algorithm from falling into local optimum.During the annealing process,the cooling parameters are oscillated,enabling the algorithm to perform an accurate search locally while speeding up the convergence.This dissertation selects a relatively high-quality initial antibody group based on the virtual machine load and the expected completion time of the task,which further improves the efficiency of the algorithm.(2)This dissertation combines the fast convergence properties of the improved simulated annealing algorithm with the efficient optimization ability of the tabu search algorithm,and proposes a fusion algorithm ISATS.The algorithm can converge quickly in the early stage,and uses the efficient optimization ability of the tabu search algorithm to enhance the search for the global optimal solution in the later stage.In addition,in the process of selecting candidate solutions,the Metropolis criterion considering the load factor is used,which can effectively expand the search range of the algorithm under the consideration of load balancing.The final experimental show that the fusion algorithm ISATS has better performance in terms of completion time and load balance.