C-RAN user association and computing resource allocation method based on deep reinforcement learning

Abstract

The invention discloses a C-RAN user association and computing resource allocation method based on deep reinforcement learning. The method comprises the following steps of 1) establishing a deep reinforcement learning neural network, and combining a signal to interference ratio (SINR) state of the neural network, a computing resource state in a baseband unit (BBU) pool and a far-end radio frequency head (RRH) cache state into a system state as input of the neural network; and 2) training the neural network according to the input system state to obtain neural network output, namely system action. and 3) using the C-RAN to perform the user association and computing resource allocation in the BBU pool according to the system action, and obtain a reward value under the system action and the system state at the next moment according to the reward function and the state transition matrix; and 4) inputting the reward value and the next system state into the neural network, repeating the abovesteps until the reward value tends to be stable so as to complete the training process, and carrying out user association and computing resource allocation in the BBU pool according to the final system action, thereby greatly reducing the service time delay, improving the service quality, and enabling the real-time service to be possible.

Description

technical field [0001] The invention belongs to the technical field of mobile communication, and in particular relates to a C-RAN user association and computing resource allocation method based on deep reinforcement learning. Background technique [0002] The 5G era is coming to support massive connectivity between humans, machines. The 5G access network is a multi-layer heterogeneous network that meets multiple scenarios, and can accommodate various wireless access technologies that have been widely used and multiple access technologies for 5G new air interfaces. C-RAN is considered as the core technology to realize these services in 5G network. C-RAN integrates the base station (Base Station, BS) infrastructure with cloud computing, and then uses optical fiber to connect to the low-latency network to form a large baseband unit (Baseband Unit, BBU) pool. This centralized baseband processing can greatly reduce the number of base stations covering the same area; wireless re...

AI Technical Summary

Problems solved by technology

However, the joint baseband data processing and cooperation between a large number of RRHs will generate a huge traffic load on the backhaul link

Images