User clustering and power distribution method and device based on deep Q network, and medium

Abstract

The invention discloses a user clustering and power distribution method and device based on a deep Q network and a medium. The method comprises steps of using a user clustering and power distributionproblem for modeling a joint optimization problem; establishing a BP neural network to realize a power distribution function in the joint optimization problem; training a BP neural network by using the training data set, testing the network and storing a BP neural network model to obtain power distribution results under different channel conditions so as to realize power distribution; modeling theuser clustering problem into a reinforcement learning task; constructing a deep Q network according to the reinforcement learning task; and after the network is trained on line, training a deep Q network according to an input state, and selecting an optimal action as an optimal clustering result to realize user clustering. According to the method, the online calculation complexity can be reduced,the user fairness is ensured to a certain extent, and the spectral efficiency of the system is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of resource allocation in a communication system, and in particular relates to a user clustering and power allocation method, device and medium based on a deep Q network. Background technique [0002] Facing the serious shortage of wireless spectrum resources and the current situation that the spectrum utilization rate of existing communication links is close to the limit, how to further improve the spectrum efficiency and system capacity to meet the requirements of large traffic, huge connections and high reliability under the full-scenario application of future wireless communication systems The demand is a key issue that research institutes in the field of wireless communication must urgently solve. Non-orthogonal and large dimensions are considered to be an effective way to improve the utilization of spectrum resources. In 2010, NTT DoCoMo of Japan proposed for the first time the power-domain Non-orthog...

AI Technical Summary

Problems solved by technology

However, the channel characteristics between different users are complex, and it is difficult for traditional methods to capture the potential relationship between users

At the same time, the solution space of the optimization problem is huge, and the nonlinear search process is inevitable

Therefore, it is difficult to obtain good user clustering and power allocation results using traditional methods, and the performance of the NOMA system is still greatly limited

The survey shows that the current research on large-scale MIMO-NOMA has not formed a systematic and comprehensive intelligent solution. The research focuses on a single perspective and the solidified deep learning network structure used has become the main reason for the limited system performance. Further technological breakthroughs are urgently needed

Images