A Research On Resource Allocation Of Fog Radio Access Network

With the increasing demand for services of mobile communication,there is an urgent need for a mobile communication system with high speed,low latency,high security,high density,wide coverage,and flexible deployment.Therefore,the researchers have proposed the C-RAN(Cloud Radio Access Network)architecture,but the front-haul in the C-RAN is constraint in capacity and delay,so the researchers have proposed the FRAN(Fog Radio Access Network)architecture with the combination of fog computing and C-RAN.There are multiple access transmission modes in F-RAN and the resource management of F-RAN can be implemented by using network slicing technology to meet diverse requirements.And this brings many resource allocation challenges,especially in combination with multiple resources.Therefore,this thesis aims at the resource allocation problem in the local distributed cooperative transmission mode in F-RAN,and summarizes the resource allocation problem into two categories.The first one is the user access resource allocation problem of multi-user multi-F-AP(Fog Access Point),and the second is a cooperative allocation problem of multiple F-AP resources in a scenario where multiple users access to the same F-AP node.The main contents and contributions of this thesis are listed as bellow.1)Aiming at the user access resource allocation problem of multi-user and multi-FAP,the user's intention to offload task to the F-AP is described as the user's evaluation of the F-AP node,and the resource allocation model based on auction theory is established.2)Three auction methods are designed on the resource allocation model based on auction theory,which are single round auction method,iterative price increase auction method and iterative price reduction auction method.Corresponding algorithms SGKP,MGKP and D-MGKP are proposed for each auction method and the differences between the three auction method in terms of total income of F-APs,total user utility and the amount of successful access users are compared.The iterative price increase auction method does not reduce the amount of successful access users.At the same time,the total income of F-APs is increased,and the user's willingness to participate in access competition is also satisfied,which is more suitable for practical application.3)Aiming at the cooperative allocation problem of multiple F-AP resources in the scenario where multiple users access to the same F-AP,a multi-F-AP node resource cooperative allocation model considering both communication resources and computing resources is established.4)In the F-APs resource allocation model,with the primary goal of minimizing the total delay of F-AP clusters,and minimizing communication overhead and load balancing as secondary goals,a heuristic algorithm(ATFGA)based on improved genetic algorithm for adaptive three fitness is proposed.The algorithm can optimize the weights for delay,load balancing and communication overhead.And we can choose a suitable scale to optimize all the three targets.The algorithm has good search ability in the global space.