Research On Resource Scheduling Algorithms Based On Cloud Computing Environment

Cloud computing is one of the hottest areas of computer research in recent years. As cloud services market and its great commercial value as emerging business models are also presented. Cloud computing data center virtualization technology resources will make various hardware and software resources abstract into virtual resources, to construct dynamic virtual resource pools, on-demand service to users. Therefore, resource scheduling becomes a key technology for cloud computing. More and more prominent commercialization characteristics of cloud computing and users increasingly diverse demand for services require service providers to pay more attention to cloud users.The different resource type of cloud computing and user preferences is varied, the target constraint of QoS will often contain more than one index requirements for users QoS. The satisfaction of QoS target constraint determines performance advantages and disadvantages of the cloud computing task scheduling strategy. To solve this problem, this paper established a cloud computing resource scheduling model. For a large number of user tasks different target constraint QoS requirements, we establish the corresponding target QoS constraints, then quantify the user’s application preferences, and construct the utility function, to make multi-target QoS constraint problem into a single target constraint problem and to use the user maximize the effectiveness as the objective function. On this basis, the paper uses improved genetic algorithm to slove the target constraint problem and make the square of the objective function as the fitness function. In genetic manipulation, we introduce an adaptive manner to maintain the diversity of the population, concentrated search for the optimal solution in a larger space, and get an optimal resource allocation policies.This paper describes the cloud simulation tool CloudSim and configure the test environment. We do simulation experiment for improving algorithm to demonstrate the feasibility and effectiveness of the proposed algorithm and to ensure the realization of the multi-dimensional QoS, minimize operating costs and optimize resource utilization. This article from the three aspects of the task completion time, cost, and the objective function value, etc. we compare our algorithm and the traditional genetic algorithm, Min-Min algorithm. The simulation results shows that the paper compared to the Min-Min algorithm is slightly worse in the completion time, but in other aspects of the algorithms are better than Min-Min algorithm and the traditional genetic algorithm, reflecting the more advantageous of our algorithm. In summary, the proposed algorithm can select the appropriate resource nodes based on user needs and get a satisfactory allocation policy.