Research On Resource Management Model And Algorithm In Fog Computing Environment

Fog computing extends cloud computing services to the edge of the network,making up for the deficiencies of cloud computing in terms of location awareness,mobility support and latency.As a new computing paradigm,the features of"using fog nodes as resource providers" of the Fog computing brings a series of new problems and challenges.Because of the resource limitation,distributivity,heterogeneity and selfishness of fog nodes,how to effectively manage resources in fog nodes has become an important research content in the field of Fog computing.In the aspect of Fog computing resource management,here are some problems that need to be solved urgently.First of all,it is difficult to effectively balance the income and cost of fog nodes when contributing resources.Secondly,cloud service providers lack effective incentive mechanisms to promote the stable and continuous contribution of resources by fog nodes.Thirdly,the quality of service of Fog computing applications is difficult to be guaranteed.This thesis studies the resource management of the Fog computing in respect of the issues above.Inspired by the working mechanism of the human nervous system,this thesis proposes a human-like neural network architecture for the Fog computing.Based on this architecture,the studies on resource contribution,incentive mechanism and resource allocation are conducted in-depth by using differential game theory,repeated game theory and improved NSGA-? algorithm.The main work and innovative results are summarized as follows:(1)The resources contribution model of Fog computing is proposed based on the differential game theory,and the optimal control scheme of resource contribution of fog nodes is constructed.Compared with the traditional static resource contribution model,the proposed model takes into account the dynamics of the fog nodes in the strategy selection process of resource contribution and the interaction of different fog nodes when contributing resources,gives the optimal control scheme for resource contribution of fog nodes in finite time domain and infinite time domain and solves the trade-offs between income and cost when contributing resources.(2)Incentive mechanism of Fog computing is proposed based on repeated game theory and optimal reward scheme for cloud service providers is constructed.Compared with the traditional incentive mechanism that lacks consideration of future factors,the proposed scheme takes into account the impact of the future revenue on the incentive strategy of cloud service provider and introduces a triggering strategy with credible threat to both players.By comparing the revenue of unlimited repeat games between cloud service providers and fog nodes,the optimal reward strategy of cloud service providers to stimulate fog nodes for the continuous and steady contribution of resources is given.(3)The resource allocation model and algorithm of Fog computing for the optimization of service delay and stability are proposed,which improve the stability of the task execution and reduce the delay of the service.Compared with the traditional Fog computing resource allocation scheme,the proposed scheme takes into consideration the significant differences in the service performance and reputation of different fog nodes.Resources are allocated in accordance with the optimization of service delay and the stability of task execution,which effectively guarantees the service quality of the Fog computing.In addition,aiming at the problem that the traditional NSGA-? algorithm is prone to produce unsatisfactory distribution of the optimal solution,the calculation method of the crowding distance is optimized,which makes the resource allocation solution set has a better distribution.Simulation results show that the proposed resource allocation algorithm has better performance in reducing service latency and improving the stability of the task execution process.