Design Of Adaptive Hierarchical Structure And Research On Key Technology In Fog Radio Access Networks

With the approach of the "Interconnected" era of mobile internet,the increasing network capacity and differentiated service performance requirements bring challenges for the traditional single-structured radio access network,and the traditional network management faces the problem of high cost and inefficiency.Considering that the fog radio access network is a high spectral efficiency and high energy efficiency networking scheme,to accommodate high capacity and differentiated service requirements,this paper designs an adaptive hierarchical structure in the fog radio access network,and studies the network intelligence and the resource management.Firstly,for the problem that the pressure of the fronthaul load in the fog radio access network is too heavy and the service performance requirements are differentiated,based on the two-level protocol architecture for the centralized unit/distributed unit and edge computing,this paper designs an adaptive hierar-chical structure,which can realize the adaptive networking of centralized and distributed,and the division and configuration of protocol stack functions under different application scenarios and transmission network conditions.Meanwhile,the deployed intelligence modules can also provide hierarchical and distributed intelligence services for the network,alleviating the pressure of the fronthaul load,meeting the needs for service personalization,data valued and network intelligence,improving the efficiency and reducing the cost of network operation and maintenance.Secondly,aiming at the problem that the intelligence service needs to realize the adaptive of centralized and distributed in the adaptive hierarchical structure,a user location intelligent prediction scheme based on the adaptive hierarchical structure is proposed.The prediction models in cloud and fog nodes are designed,and the adaptive cloud and fog joint prediction strategy is also proposed.The performance of the proposed scheme is verified by experiments,and the adaptive application gain of network intelligence in the structure is given.Finally,for the optimization problem of minimizing the total calculation delay of online prediction under resource constraints in the adaptive hierar-chical structure,the paper designs a joint optimization scheme of online prediction model deployment and task assignment.In order to minimize the total calculation delay of online prediction in the network,low complexity solution methods based on greedy algorithm and heuristic algorithm are proposed to solve the above problem,respectively.By comparing with the backtracking algorithm which can obtain the global optimal solution,the optimal algorithm under different network conditions is obtained,which improves the adaptive ability of the network.Based on the adaptive hierarchical structure in the fog radio access network designed in this paper,combined with the intelligence service scheme and the online prediction model deployment and task assignment joint optimization strategy in the structure,the wireless network can realize the adaptive networking of centralized and distributed and intelligence prediction under different application scenarios and transmission network conditions,and minimize the total calculation delay of online prediction,alleviating the pressure of the network load,meeting the need of differentiated service performances of users and realizing network intelligence.