Research On The Allocation Mechanism Of IoT Edge Computing Resources Based On Game Theory

With the popularity of Io T and 5G technology,edge computing(EC)has been developing rapidly in the past few years.This computing model deploys computing capabilities at the edge of the network close to users,which can improve user experience,reduce network latency,save user resources,and bring more convenient and efficient services to people.However,as users’ demand for data volume and computing power grows,EC also faces many challenges.Among them,how to effectively solve the problem of resource sharing and computational offloading in the Io T edge environment through collaborative computing is a current research hotspot and difficulty.To this end,coalition game technology can be used to analyze the strategic interactions among multiple cooperative entities,and the network entities in edge computing can form a reasonable coalition structure,and then effectively solve the resource competition and benefit maximization problems existing in edge computing.Based on the above background,targeting both end-to-end and end-to-edge application scenarios,this paper proposes a coalition collaborative offloading scheme based on device task matching value and a coalition game-based offloading decision and revenue allocation scheme for offloading decision,resource sharing and cooperative revenue allocation in edge computing,respectively,and the main work is as follows.Firstly,in end-to-end edge network environments,due to the imbalance in computing resources and demands of terminal devices,a large amount of idle resources are often unable to be fully utilized.In order to fully tap the idle computing resources of Edge device and effectively alleviate the resource shortage in edge environments,this paper proposes an alliance game unloading scheme based on task matching values between devices.Initially,each terminal device forms an initial coalition structure as a single device coalition.Then,each device searches for a set of idle devices that can be offloaded within its communication range,using the task matching values among devices in the coalition as the condition.Based on the rules set in the scheme,coalition switching and merging operations are performed.After repeated iterations,the coalition converges to a stable state.The simulation experimental results show that compared with existing research results,this scheme can effectively reduce the unloading failure rate of tasks with different parameter types and significantly improve user utility.Secondly,unlike in end-to-end task offloading scenarios,in order to alleviate terminal computing pressure and make reasonable use of edge service computing resources in end-to-edge offloading scenarios,this paper proposes a cooperative offloading scheme based on alliance games.MEC servers share their computing resources to process the computational tasks offloaded by UDs,thereby improving the task completion rate.The price mechanism is used to incentivize both parties to cooperate.UDs pay a service fee to the MECs in the coalition while offloading their tasks.The stability of the large coalition formation is theoretically analyzed,and the super-additivity of coalition benefits is demonstrated.Meanwhile,to address the issue of fair distribution of coalition benefits,the concept of nucleolus and Shapley value in coalition game is introduced,and two different allocation schemes are realized.Simulation results show that the grand coalition structure proposed in this scheme has the highest overall benefit and also the two different allocation schemes achieve a more equitable distribution of benefits,both in terms of overall utility and individual utility,to obtain the optimal benefit.