Research On Resource Allocation Problem Of Electronic Health Network Assisted By Edge Computin

The e-health network can not be limited by time and space,and provide medical care services anytime and anywhere,which has great application potential and research value.However,the communication and computing resources of the e-health network are limited,and the efficient operation of the e-health network depends on the reasonable allocation of resources.Aiming at optimizing the resource allocation in the multi-level e-health network composed of the end users,the edge servers and the cloud server,the paper studies from the following three aspects:communication resource allocation of the end user layer,the computing resource allocation of the edge layer and cloud layer,CPU frequency adjustment and communication resource allocation of the edge layer.The main work of the paper is as follows:1)The communication resource allocation problem at the end user layer is studied.In the scenario of ‘multiple end users–single edge server’,aiming at the spectrum scarcity caused by the increase of the number of terminal devices,a dynamic spectrum access scheme for end users based on cognitive radio technology is proposed.Considering the complex communication model of inaccurate channel sensing,spectrum resource availability and signal interference,the dynamic spectrum access problem of multi terminal devices for optimal throughput is constructed.Then we designs the dynamic spectrum access strategy based on the double Q network algorithm,which effectively optimizes the channel allocation in the user layer of the e-health network.Simulation results show that the proposed scheme can improve the throughput,data transmission success rate and collision rate of the network in various typical spectrum environments.(2)The computational resource allocation of edge layer and cloud layer are studied.In the scenario of ‘multiple end users–multiple edge servers’,the task offloading model of the end users and the resource transaction model in the edge servers are constructed respectively for the limited computing resources of end users and the unfair allocation of mining resources of edge server.And an improved reputation scoring mechanism is established to evaluate the computing contribution of the edge servers.Based on this,the paper further proposes a two-stage task offloading and hash resource allocation mechanism to realize task offloading decision and hash resource allocation.Simulation results show that this mechanism can significantly reduce the delay and energy consumption of the system and prolong system life cycle.(3)The CPU frequency adjustment and communication resource allocation in edge layer are studied.In the scenario of ‘multiple end users–multiple edge servers’,the CPU frequency adjustment and channel access strategy for the edge server is proposed in the paper in order to improve the energy efficiency of federated learning.The cognitive radio technology is used to enrich spectrum resources.The channel model,delay model and energy consumption model are established.The resource optimization problem is solved by deep reinforcement learning,so as to maximize the number of federated learning rounds of the edge server.Simulation results show that the proposed strategy can effectively reduce the energy consumption of the edge server,reduce the number of collisions with the primary user in the channel,and improve the practicability of federated learning in the e-health network.