Research On Joint Optimization Technology For Energy Efficiency Transmission And Computing Resources

Due to the rapid growth of multimedia services and the large increase in carbon dioxide emissions,the application of renewable energy(green energy)to wireless communications is imminent.In many cases,communication systems are primarily powered by renewables.For example,in remote mountainous areas,grid deployment costs are high and electricity prices rise,at which point renewable energy is used as the primary source of energy.Some technologies have proven to be effective ways to achieve green wireless access,such as energy capture,multicast,and edge computing.Some of the current communication technologies have proven to be more suitable for access to renewable energy sources such as content transmission,multicast,cloud computing,mobile edge computing,and the like.In addition,technological advances in smartphones,laptops and tablets have enabled demanding services and applications.But these services and applications mean lower latency,higher computational speed,higher stability,and more.In view of the above problems,this paper mainly studies the joint optimization technology for energy-efficient transmission and computing resources.This includes the issue of content transfer and the removal of mobile edge calculations.First,we solved the problem of green energy capture and mission arrival mismatch.Starting with issues related to content broadcasting and caching,small base stations powered by fully green energy.Considering that the arrival of green energy is random and difficult to predict,energy waste or shortage will occur when energy and mission arrivals do not match.Therefore,we aim to improve the utilization of green energy and propose a content push based on Q learning.Resource management algorithm.Q learning is a model-free,enhanced learning technique that finds the best action selection strategy in MDP problems.The SBS selects the action according to the Boltzmann strategy and then iteratively updates the Q table to get the best action in each state.After the simulation,the algorithm improves the utilization rate of green energy,and obtains the law of SBS decision,which proves the validity of the model.Second,we propose an efficient task-migration strategy based on reinforcement learning,which can understand the dynamic workload offload and the state of the edge server in real time,and give the best strategy.It also addresses the challenge of incorporating renewable energy into mobile edge computing.The algorithm is used by the IoT device to select the MEC device,and determines the offload ratio based on the current battery power,the previous radio bandwidth of each MEC device,and the predicted amount of collected energy.The simulation results show that the method effectively reduces network delay,packet loss rate and improves energy utilization efficiency.