Research On Resource Allocation Method Based On Auction Mechanism In Edge Computing Environment

With the continuous development of the Internet and mobile applications,the mobile services available to users are gradually enriched and improved.The richness of service functions and the complexity of mobile applications make the computingintensive and delay-sensitive application characteristics of mobile services more and more obvious.In recent years,computing offload and other technologies have been proposed to enable mobile devices to offload their tasks to the remote resource-rich cloud,to overcome the resource constraints of mobile devices.However,the current cloud-based computing model still has a deficiency of high latency and can not meet the quality of service requirements of mobile users.Driven by the above applications,edge computing technology is gradually developed.By docking mobile services such as computing power,communication capacity,storage capacity,and other mobile services with mobile users,edge service providers can provide corresponding edge computing resources to application service providers.Application service providers can deploy applications to edge servers to allow users to obtain services with low latency.Resource allocation in edge computing environment involves more and more application service providers and edge service providers.The benefit game between the edge service providers and the application service providers makes the resource allocation in edge computing environment become a technical and application problem that needs to be solved urgently.To solve the above problems of resource allocation in edge computing environment,this thesis makes targeted research on the framework and methods of resource allocation in edge computing environment.Specifically,the main research work of this thesis is as follows:(1)To meet all kinds of resources' requirements in different application scenarios,this thesis analyzes the application needs of application service providers and the profit requirements of edge service providers.A resource allocation framework based on auction is proposed according to the application characteristics of edge computing.The resource allocation framework is divided into three layers from top to bottom,namely,service layer,platform layer,and resource layer.The service layer is the resource requirements of the application service provider in different application scenarios.The platform layer involves various modules of auction,including user registration,authentication,specific auction methods,payment,and other functions.These are the core components of the framework.The resource layer is that the layer where edge service providers can offer all the resources.(2)Aiming at the application requirements of mobile service providers with single-type resource requirements(e.g.,video caching),an edge computing single-type resource allocation method is proposed based on an auction mechanism.The method consists of an algorithm for determining the auction winners and the transaction price and a resource allocation algorithm.First of all,the effective price is calculated according to the bidding information of the active auction participants and the supply and demand information.Then the auction winners and the final transaction price are determined according to the effective price.Among the winning auction participants,the final allocation of resources is obtained according to the constraints of supply and demand.Through the analysis of the proposed algorithm,this method can meet the characteristics of personal rationality,incentive compatibility,and weak budget balance.The experimental analysis of different indicators shows that the proposed method has ideal results in resource allocation,utilization,utility,and so on.(3)Aiming at the application requirements of mobile service providers with multitype resource requirements(e.g.,augmented reality),a multi-type resource allocation method is proposed based on auction mechanism in edge computing environment.Specifically,we first discuss the combinatorial double auction model for multi-type resources in this method.To facilitate the solution of the model,we transform the model and prove that the transformed model is essentially a multi-dimensional knapsack problem at the beginning.After that,the genetic algorithm is introduced to solve the multi-dimensional knapsack problem.Finally,the auction results are obtained by the resource pricing and allocation algorithm,that is,the auction transaction allocation matrix and the transaction price matrix.A series of simulation experiments are carried out to verify the resource utilization and system utility of the proposed algorithm.