Research On High Utility Resource Optimization In Fog-Computing Radio Access Network

In recent years,with the continuous emergence of technologies such as holography,digital twins,and telemedicine,as well as the intelligent development of global mobile terminals,the current mobile data traffic is growing in exabytes in wireless networks.Thus,the demand for caching and computing resources at network edge is also increasing.To provide high-quality experience to users,the main challenge faced by existing wireless networks has become the contradiction between the growing number of new and complex services and the limited resources of mobile networks.Fog-computing radio access network(F-RAN)is a new type of distributed network architecture that can perform operations such as storage,communication,control,configuration,measurement and management at edge nodes close to Fog-computing user equipment(F-UE),which makes F-RAN very suitable for applications featuring extremely low latency,dynamic real-time data,and large-scale distribution.Efficient utilization of communication,caching and computing resources is the key to fully exploiting the performance potential of F-RAN.At the same time,to further improve the effectiveness of F-RAN,it is crucial to establish a comprehensive F-RAN performance evaluation framework.However,traditional network communication performance indicators such as link capacity,energy efficiency,latency,and bit error rate cannot effectively characterize the performance of F-RAN.Therefore,taking into account communication,caching and computing performance,this thesis establishes utility functions related to energy efficiency,system throughput,latency and energy consumption and studies the efficient resource management of F-RAN.The main work and innovations are summarized as follows:1.Research on fronthaul bandwidth allocation and power control scheme in mm Wave based F-RAN.To meet the wireless communication requirements of low latency,huge bandwidth and hyper-connectivity,this paper combines millimeter wave with F-RAN,and adopts in-band fronthaul of the system to relieve the wireless spectrum tension and ensure a higher data transmission rate.In the proposed mm Wave based F-RAN,the limited link of high deployment cost is replaced by the flexible wireless fronthaul to connect the Fog-computing access point(F-AP)and the remote cloud center.At the same time,the bandwidth of the entire system is shared between the wireless access link and the fronthaul with the effective spectrum resources fully utilized through in-band fronthaul.Taking into consideration fronthaul constraints,edge caching,maximum transmit power and Quality of Service(QoS)constraints,a utility function for energy efficiency of FUE is established,and bandwidth allocation and power control methods in mmFRAN are studied to maximize the utility.Its utility maximization is a non-convex nonlinear optimization problem,which can be proved to have a global optimal solution.In order to reduce the solution complexity,the original problem is decoupled into a bandwidth allocation subproblem and a power control subproblem to be solved separately.Then,the optimal resource allocation algorithm can be obtained through alternating iterations of two subproblems,and a low complexity suboptimal algorithm is proposed.Finally,the effectiveness of the proposed algorithm is proved by simulation experiments.2.Research on joint computation offloading and resource allocation scheme in F-RAN.To make full use of the computing resources on the edge side of F-RAN so as to meet the demands of increasingly complex terminal applications,this paper adopts the computing offloading technology.The computing task can be offloaded from the F-UE to the F-AP with higher computing ability by wireless transmission,and the computing resources at the F-AP can be used to improve the network edge response capability of more complex services.When establishing the optimization problem,the latency sensitivity of the F-UE and the limited power reserve are also considered,and the latency trade-off factor and the energy consumption trade-off factor are introduced to ensure the trade-off balance between the latency performance and the energy consumption overhead.To better show the reward of the computation offloading,the offloading utility function is defined,and the objective problem of maximizing the total utility of all F-UEs in the system is established.Since the utility maximization problem is an integer mixed optimization problem,considering the complexity of the solution,the original problem is decoupled into a computational offloading decision subproblem with fixed power and computing resources and a resource allocation subproblem with a fixed offloading strategy.The computing offloading strategy is obtained by the Hungarian algorithm,and then the closed-form solutions of power control and computing resource allocation are obtained by convex optimization.Then through alternate iterations,a joint computing offloading and resource allocation algorithm for F-RAN is obtained.Finally,through simulation verification,it is verified that the proposed algorithm can achieve a certain offloading gain,which significantly improves the ability of F-UE to deal with complex applications at network edge.3.Research on dynamic resource management scheme in Non-orthogonal multiple access(NOMA)based F-RAN.Resource optimization in F-RAN.To cope with the exponential growth of wireless network access requirements,this paper combines NOMA technology with F-RAN to increase the access scale so that the system can obtain higher spectral efficiency.For NOMA based F-RAN,considering QoS constraints,maximum transmit power constraints,system stability and other conditions,a stochastic optimization model is established,and a resource allocation algorithm that balances link congestion and network utility is proposed.The subchannel allocation subproblem is first solved by using matching game theory,in which the channel allocation is established as a dynamic matching process between multiple F-UEs and multiple F-APs to obtain the subchannel allocation strategy.Then,according to Lyapunov theory,Lyapunov minimum drift term and penalty are introduced,and the average utility of F-RAN is optimized through power allocation when system stability is ensured.The resulting resource management strategy can affect the balance between link congestion and average utility value by selecting different control parameters.Through simulation verification,the effectiveness of the proposed dynamic resource optimization algorithm is proved from the perspective of average utility and queue stability.