Research On D2D Communication Resource Allocation Mechanism Based On Cellular Network

Device-to-Device(D2D)communication is a widely researched and promising technology which can effectively increase the system capacity of cellular communication networks while reducing the load.It is also available in 5G which will be put into commercial use.It also has good practical value and development space.Since D2D communication usually reuses traditional cellular communication networks,interference between users is inevitable.The main methods of controlling interference include resource allocation and power control.Therefore,efficient energy management has become the main research direction of D2D communication.This paper proposes a user resource allocation algorithm for D2D communication,which maximizes system throughput while coordinating interference.Due to the uncertainty of the wireless channel state and state transition possibilities,it is considered that the Q learning algorithm in the Reinforcement Learning(RL)algorithm can be used to solve this problem.In the solution presented herein,each D2D pair is treated as a separate agent,making corresponding decisions on resource selection based on locally observed channel states.In order to solve the sequence decision problem in multi-user system,a multi-agent RL algorithm is further proposed.Under the premise that the D2D pair does not have any information about the availability and quality of the resource blocks to be allocated,the problem can be modeled as a random non-cooperative game.Each agent can be regarded as a participant.Through the joint decision of all participants to achieve global optimization,a multi-agent Q learning algorithm based on game theory is proposed.In addition,this paper starts from the perspective of adjusting the transmission power of each D2D user to maximize the system throughput while adjusting the interference.Similarly,the problem can be modeled as a reinforcement learning algorithm.In order to improve efficiency and reduce convergence time,firstly,the fuzzy clustering algorithm is used to group D2D users,so that there is only small interference between users in the same group,and then each group is regarded as an agent in the Q learning algorithm,thereby establishing A multi-agent Q learning algorithm model based on fuzzy clustering algorithm.For the above two algorithms,the corresponding simulations are carried out,and the performance of the user can be improved while ensuring the communication quality of the user.The superiority of the proposed algorithms are verified.