Strategy-Proof Auction Mechanism For Resource Allocation In Mobile Edge Computing

The resource allocation problem based on the auction mechanism is one of the hot issues in the current mobile edge computing research.Using the auction mechanism to allocate computing resources can significantly improve the mobile edge computing service provider's ability to utilize idle resources and increase revenue.However,the present existing research on the resource allocation auction mechanism in mobile edge computing still has many shortcomings.Firstly,the existing mobile edge computing resource allocation auction mechanisms don't consider the strategy-proofness.Secondly,the existing auction mechanisms cannot allocate resources of users' multi-computing tasks to different resource servers,which is not conducive to improving the social welfare and resource utilization of resource providers.Thirdly,the existing auction mechanisms fail to consider the allocation of bandwidth for establishing connection between user mobile devices and mobile edge computing.Fourthly,the existing auction mechanisms don't support online resource allocation to multiple servers.In response to the above problems,this thesis studies the resource allocation problem in the mobile edge computing environment from the perspectives of offline auction mechanism and online auction mechanism.The specific contents are as follows:(1)For the perspective of offline resource allocation in the mobile edge computing environment,this thesis considers the scenario where offline resource allocation includes multiple users and multiple servers,each user has multiple computing tasks,and supports bandwidth access.Then an offline strategy-proof auction mechanism is proposed.This mechanism uses a depth-first search strategy to allocate user's multi-tasks on multiple servers,and uses resource density and task adaptability on the server to allocate resources to ensure the effect of resource allocation,while the amount of calculation is reduced.This thesis also proves that the mechanism is strategy-proof,and analyzes the approximate ratio and time complexity.The experimental results show that the offline strategy-proof auction mechanism proposed in this thesis can obtain the approximate optimal solution of resource allocation in a short time,and can significantly improve social welfare and resource utilization compared with other resource allocation algorithms.(2)For the perspective of online resource allocation in the mobile edge computing environment,this thesis considers the scenario where online resource allocation includes multiple users and multiple servers,and supports bandwidth access,and introduces time as a parameter for mathematical modeling.Then an online auction mechanism is proposed.This mechanism uses resource density and task adaptability on the server to allocate resources to reduce the amount of calculation while ensuring the effect of resource allocation.This thesis also proves that the mechanism is strategy-proof,and analyzes the time complexity.The experimental results show that the online strategy-proof auction mechanism proposed in this thesis can obtain the approximate optimal solution of resource allocation in a short time,and can significantly improve social welfare and resource utilization compared with other resource allocation algorithms.