Reinforcement Learning Guided by Double Replay Memory

Figure 1

Simulation results via CartPole to compare to DQN and PER.