Residual-gradient-based neural reinforcement learning for the optimal control of an acrobot.
Xin XuHan-Gen HePublished in: ISIC (2002)
Keyphrases
- optimal control
- reinforcement learning
- control problems
- dynamic programming
- model free
- fitted q iteration
- network architecture
- rl algorithms
- class of nonlinear systems
- function approximation
- optimal control problems
- infinite horizon
- feedback control
- neural network
- risk sensitive
- actor critic
- brownian motion
- learning algorithm
- control strategy
- reinforcement learning algorithms
- control law
- state space
- linear quadratic
- average cost
- optimal policy
- machine learning