Abstract
We present a strategy for learning optimal control. The approach uses functional-link neural network implementations which have several beneficial properties giving advantages over the more common generalized delta rule implementations. The learning task is decomposed into three parts: identification and monitoring, one-step-ahead control generation, and control path optimization. Each of these parts is accomplished with its own functional-link net and these are coordinated to provide the real-time learning of the optimal control path.
Original language | English (US) |
---|---|
Pages (from-to) | 149-164 |
Number of pages | 16 |
Journal | Neurocomputing |
Volume | 9 |
Issue number | 2 |
DOIs | |
State | Published - Oct 1995 |
Externally published | Yes |
Keywords
- Functional-link net
- Neural net control
- Optimal control
- Real-time learning
ASJC Scopus subject areas
- Computer Science Applications
- Cognitive Neuroscience
- Artificial Intelligence