Fig. 8From: Reinforcement learning-based hybrid spectrum resource allocation scheme for the high load of URLLC servicesDelay performance comparison of different schemes. The delay comparison of the RL-based, Greedy, Random, SFF, and Tetris schemes at varying loads is shown in figure. Among them, the RL-based scheme selected 500 times, 1000 times, and 10,000 times training data, respectively. The x-axis is the loads, and the y-axis is the average delay. We record the delay performance of each scheme with 9 sampling points from load 0.11 to 1.28Back to article page