Systems Science Friday Noon Seminar Series

Files

Download

Download (2.1 MB)

Date

4-8-2011

Abstract

Humans have the ability to make use of experience while performing system identification and selecting control actions for changing situations. In contrast to current technological implementations that slow down as more knowledge is stored, as more experience is gained, human processing speeds up and has enhanced effectiveness. An emerging experience-based (“higher level”) approach promises to endow our technology with enhanced efficiency and effectiveness.

The notions of context and context discernment are important to understanding this human ability. These are defined as appropriate to controls and system-identification. Some general background on controls, Dynamic Programming, and Adaptive Critic leading to Adaptive Dynamic Programming (ADP) will be provided.

The higher-level application of Adaptive Dynamic Programming (ADP) is described, wherein ADP is employed to develop on-line algorithms that respond to changes in context by efficiently and effectively selecting designs from a repository of existing controller solutions– in contrast to the usual application of ADP that focuses on designing controllers directly. In this way, the ADP is said to be applied up a level from typical application.

Key components of the approach include the notions of context, context discernment, and experience. These apply to applications in control and also to system identification.

Details of the approach and its rationale will be described, including examples and recent developments of the underlying ideas.

Biographical Information

George G. Lendaris is Professor of Systems Science and Electrical & Computer Engineering at Portland State University.

Subjects

Dynamic programming, Reinforcement learning, Adaptive control systems -- Mathematical models, Computational intelligence, System identification

Disciplines

Computer Sciences | Theory and Algorithms

Persistent Identifier

https://archives.pdx.edu/ds/psu/31163

Higher-level Application of Adaptive Dynamic Programming/reinforcement Learning – A Next phase for Controls and System Identification?

Share

COinS