Research On Time-Varying Multidimensional Resource Allocation Based On Auction Mechanisms In Edge Computing

The traditional cloud computing has the limitations in terms of working principle and technological development,such as delay,network bandwidth and security issues,etc.Therefore,it is difficult to meet the network requirements of the 5G era.It is a general trend for computing centers to sink from the cloud to the edge,so the edge computing is emerging.Edge computing has attracted the attention and application of resource providers and users because of its advantages such as low latency,low bandwidth costs,and protecting of security and privacy.Therefore,more and more resource providers deploy data generated by users to the edge.As the number of network edge devices increases exponentially,the data generated by them increases rapidly,which puts forward higher requirements for low latency.Therefore,a reasonable resource allocation scheme is particularly important for both resource providers and users.In the current field of resource allocation in edge computing,many resource allocation problems can be transformed into a knapsack problem.For this reason,resource providers have introduced auction mechanisms in edge computing to obtain more profit,allowing idle resources to be sold at dynamic prices.Furthermore,the actual resource requirements of users may change with time,and it will result in low utility of resource providers and even cause a waste of resources if the resource requirements of users are satisfied with a fixed amount of resource during the usage time.Therefore,timevarying multidimensional resource allocation is a very important research direction in the field of edge computing currently.In this thesis,we mainly used combinatorial auction mechanism and online auction mechanism to solve the time-varying multidimensional resource allocation problem in edge computing.And we established two resource allocation models and designed the resource allocation algorithms correspondingly,which aims at maximizing the total utility of resource providers.The main contents of this thesis are as follows:First of all,for the offline resource allocation in the time-varying multidimensional resource allocation problem in edge computing,users arrive at the same time.We applied combinatorial auction mechanism to establish a time-varying multidimensional resource allocation model for edge computing,which aims at maximizing the utility of the fog resource pool.In order to achieve a reasonable resource allocation,we proposed a combinatorial auction mechanism-based resource allocation algorithm and discussed the specific implementation steps of this algorithm.Secondly,for the online resource allocation in the time-varying multidimensional resource allocation problem in edge computing,users arrive dynamically.We established a time-varying multidimensional resource allocation model for edge computing based on online auction mechanism,which aims at maximizing the total utility of the cloud resource provider and fog resource pool.In addition,we proposed an online auction mechanismbased resource allocation algorithm to achieve a reasonable resource allocation and discussed the specific implementation steps of this algorithm.Finally,we conducted two numerical simulation experiments for the time-varying multidimensional resource allocation based on combinatorial auction mechanism and online auction mechanism,which verify the effectiveness of the proposed resource allocation schemes.And we gave the enlightenment of management,which is of great significance for the time-varying multidimensional resource allocation in the field of edge computing.