Research On Resource Allocation In Mobile Edge Computing Based On Game Theory

With the rapid development and convergence of the Mobile Internet and Internet of Things(Io T),the number of mobile terminals and the cloud-oriented applications have been increased.However,due to high latency,low coverage,and poor scalability,the architecture of the Cloud Computing and Mobile Cloud Computing cannot meet the computation requirements of ultra-low latency and ultra-high data capacity.And the emergence of compute-intensive and delay-sensitive applications(such as Augmented Reality,Virtual Reality and Driverless)are driving task computing and data storage to the edge of Internet.As a new computing paradigm,Mobile Edge Computing(MEC)effectively improving the service experience of users,and drawing a lot of studies and researches.However,in the distributed MEC environment,heterogeneous task usually have different requirements in computational latency.At the same time,due to the complicated and time-varying characteristics of date,there are serious challenges for efficient task offloading and resource allocation.In the time-varying MEC environment,this thesis modeled a diversified task offloading model by analyzing the heterogeneous task requirements in offloading latency and computation resource.Furthermore,in order to realize the flexible control of task offloading and the on-demand allocation of computation resources,this thesis design a dynamic on-demand quote method for computation resources thought making a tradeoff between offloading revenue and latency by invoking the Buyer-Seller game and Lyapunov optimization theory.Finally,the ondemand resource allocation algorithm for heterogeneous tasks is proposed.Simulation results demonstrate that the proposed algorithm can meet the differential computation requirements of heterogeneous tasks and reduce the average latency of the MEC system while ensuring network stability.In addition,due to the suddenness of task requests of mobile users and the uncertainty of the MEC servers queue time,inaccurate predictions of the MEC servers queue time will cause computing failed to delay-sensitive tasks.To address this problem,based on the on-demand resource allocation algorithm for heterogeneous tasks algorithm of above,this thesis further consider the uncertainty of the MEC servers queuing time,and by invoking the Buyer-Seller game and multi-stage stochastic programming theory to model a multi-stage stochastic programming based users and MEC servers distributed game model,respectively.And then,a resource allocation algorithm based on stochastic game is proposed in an uncertain environment.Furthermore,the existence of Stackelberg Equilibrium(SE)is proved.Finally,simulations verify that the proposed algorithm can effectively improve the success rate of task offloading while maximizing the long-term rewards.