| In recent years, Mobile Internet business develops rapidly. Compared with PC and traditionalInternet, mobile cloud computing has characteristics such as “thin”, network bandwidth anddelay requirement. This makes effective resource management become the key problem of themobile cloud computing.In this paper, cloud computing and mobile cloud computing are studied deeply. First of all, itanalysis the resources management requirements of mobile cloud computing. Secondly, amanagement model based on multi-layer mapping is designed for the virtual resourcemanagement platform. Then, a resource scheduling algorithm based on the cellular particleswarm is proposed. The paper also proposes a resource allocation prediction algorithm based onwavelet decomposition and support vector machine. Finally, it verifies the correctness andsuperiority of the two algorithms through experiments. It also tests the effect of the virtualresource management model.The mainly innovations of this essay are as follow:(1) It proposes the hierarchical virtual resource management model. To the schedulinggranularity and flexibility, it needs a service-resource mapping model. It also gives aresource scheduling scheme to achieve fine-grained user requests.(2) It proposes the virtual resource scheduling algorithm based on cellular particle swarm.By analyzing the characteristics of virtual resource scheduling problems, it builds areasonable model for multi-objective optimization. Combined with particle swarmoptimization algorithm and cellular automata, a reasonable and efficient resourcescheduling algorithm is designed.(3) It proposes the resource allocation prediction algorithm based on wavelet decompositionand support vector machine. According to the DB2wavelet decomposition and SVM, itproposes a prediction algorithm based on wavelet analysis, SVM modeling and waveletreconstruction. Through the study of historical resource data, the algorithm can predictthe resource allocation accurately in the next time node. |
PDF Full Text Request