ETRI-Knowledge Sharing Plaform

Training Data Real-World Robust Policy Off-Policy Numerical Simulation(Trnsys16) State-Of-The-Art Structure-Aware Additional Information Swing-Up Control Markov Decision Process Data-Driven Virtual Experience Need For Large Amounts Of Data Model-Free Performance Degradation Training Process Actor-Critic Algorithm Control Task Control Capability Deep Reinforcement Learning Robotics Applications Triple Inverted Pendulum Computational Efficiency Reward Function Real-Time Control Convergence Speed Experience Replay Learning Agent Virtual Samples And Wear Network Learning Learning Process