Approximate neural optimal control with reinforcement learning for a torsional pendulum device.
Ding WangJunfei QiaoPublished in: Neural Networks (2019)
Keyphrases
- optimal control
- reinforcement learning
- control problems
- dynamic programming
- control strategy
- feedback control
- network architecture
- infinite horizon
- class of nonlinear systems
- optimal control problems
- risk sensitive
- control law
- fitted q iteration
- neural network
- machine learning
- brownian motion
- function approximation
- rl algorithms
- model free
- state space
- average cost
- reinforcement learning algorithms
- inverted pendulum
- policy evaluation
- actor critic
- learning algorithm
- markov decision processes