Offline Supervised Learning V.S. Online Direct Policy Optimization: A Comparative Study and A Unified Training Paradigm for Neural Network-Based Optimal Feedback Control.
Yue ZhaoJiequn HanPublished in: CoRR (2022)
Keyphrases
- feedback control
- supervised learning
- optimal control
- online learning
- training set
- closed loop
- adaptive control
- real time
- reinforcement learning
- optimal design
- dynamic programming
- training samples
- infinite horizon
- open loop
- semi supervised
- asymptotically optimal
- neural network
- unsupervised learning
- optimization algorithm
- training examples
- learning algorithm
- training data
- expected cost
- allocation policy
- fractional order
- optimal solution
- feedback controller
- active learning
- sufficient conditions
- decision trees
- control policy
- inverted pendulum
- multiple instance learning
- dynamic environments
- supervised training
- labeled data