Research On Cloud Service Composition And Scheduling For Multi-objective Optimization

With the further development of the mobile Internet and mobile applications,cloud service with single function can’t meet the requirement of large-scale and complex application demands.So cloud services composition to meet needs of multi-function are attracting attention.When cloud service management platform dealing with the demands of the batch service requests for using limited cloud services,it also needs to optimize the scheduling strategies and improve the efficiency of the cloud service requests execution.So the research on cloud service composition and scheduling has certain theoretical significance and practical application value.This paper mainly finished the following several aspects of the research:(1) The service quality attributes of cloud services were described,and the mathematical expressions were given.Then,the optimal model about cloud service composition selection was put forward,which was based on global Qo S evaluation combined with local Qo S evaluation.(2) A cloud service scheduling model for batch service requests was established.On the basis of the scheduling model,the structure and process of cloud service composition and scheduling based on two stage optimization were described in this paper.At the same time,we determined the constraint rules of multi-objective optimization.(3) The dynamic acceleration parameters,hybrid strategy and two-order oscillating equation were the popular way to modify Particle Swarms Optimization(PSO),which were introduced to improve traditional particle swarm optimization algorithm in this paper.The Modified Particle Swarms Optimization(MPSO) algorithm was used to solve this mulit-objective problem proposed in this paper.(4) Finally, this article has carried on simulation experiment about the cloud service composition and scheduling problems with different problem scales.Experimental results showed that compared with other traditional methods, the proposed method based on MPSO to solve cloud service composition and scheduling problems proposed in this paper can improve the quality of the solution to a certain extent.The computation efficiency of algorithm proposed is high.we also did the instance simulation for cloud service scheduling model and algorithm proposed in this paper,which was proved that it had certain actual application value.