Research On Scheduling Technology Of Cloud Service Resource Based On Game Model

Cloud computing is a distributed computing model for service oriented architecture and his characteristics are open,intelligent and personalized.The prominent unified scheduling ability for services such as scattered information management,distributed computing,storage,software and hardware resources can meet the needs of information war to fight against the joint command to a certain extent.Due to the increasingly large scale of cloud computing,cloud resources have the characteristics of the quantitative and dynamic.The key problem to be solved is how to design reasonably cloud computing service resource scheduling mechanism.Based on building the cloud service resource scheduling framework,this paper analyzes the user needs,and studies the service resource scheduling technology and virtual resource scheduling technology,so as to improve the efficiency and fairness of cloud service resource scheduling.Main contents and contributions are as follows:1.Aiming at the characteristics of cloud computing,design the cloud service resource scheduling framework.The framework mainly includes two parts as service resource scheduling and virtual resources scheduling.The idea of using the game theory to study the cloud service resource scheduling technology is put forward,and the key technologies involved in the two parts are analyzed.2.Aiming at the situation of users requesting cloud service resources at the same time,design the cloud service resource scheduling method based on the static game theory.The traditional resource scheduling only considers unilateral scheduling efficiency and easily ignores the resource allocation fairness.The new method introduces static Bayesian game theory and analyses the problems of the user's competition in resource scheduling and resource allocation from the perspective of users,so as to make the resource scheduling scheme both efficient and fair.3.For multi-user successively request cloud service resources,this paper designs dynamic game model for reference service resource scheduling method.Aiming at problems existing in the traditional resource scheduling technology,applying dynamic game theory to analyze the formalization description and set up the process of resource scheduling game model,the method solves the earnings game equilibrium,designs resource scheduling algorithm and implement quantitative calculation on the basis of quantitative calculation of earnings.4.Aiming at the characteristics and requirements of virtual resource scheduling in cloud service,a virtual resource scheduling method based on improved multi-objective particle swarm optimization algorithm is designed.The resource scheduling problem is converted to the constraint condition multi-objective optimization problem by this method.This method speeds up the algorithm convergence speed with using Sigma method.When population diversity loses during the execution of the algorithm,this method introduces chaos disturbance mechanism in the solution space of the sparse particles near to generate new particles and guides the population space evolution to advantage.The algorithm avoids into the local optimal.Compared with the traditional algorithm,the improved multi-objective particle swarm optimization algorithm has faster speed and global optimization ability,which is of great significance to improve the efficiency of cloud resource scheduling.