Deep Reinforcement Learning Based Resource Allocation for URLLC User-Centric Network

In this paper, we solve the resource allocation problem by deep reinforcement learning (DRL) for diverse ultra-reliable low-latency communication (URLLC) services under the user-centric downlink transmission. Firstly, to meet the constraint of reliability, we model the channel decoding error rate by using the finite blocklength coding (FBC) according to the short packet characteristics of URLLC services. Then, we model the queue of different URLLC services in the temporal dimension to describe the delay violation problem. Furthermore, we adopt the DRL scheme that transforms the maximizing system availability and transmission efficiency problem into maximizing system reward problems. Simulation results show that the proposed algorithm achieves superior availability for diverse URLLC services compared with the baselines.