| | |
Summary: THESIS
CONTINUOUS REINFORCEMENT LEARNING FOR FEEDBACK CONTROL SYSTEMS
Submitted by
Jilin Tu
Computer Science Department
In partial ful llment of the requirements
for the Degree of Master of Science
Colorado State University
Fort Collins, Colorado
Summer 2001
ABSTRACT OF THESIS
CONTINUOUS REINFORCEMENT LEARNING FOR FEEDBACK CONTROL SYSTEMS
To achieve the best control performance, feedback control techniques require a very accurate
mathematical model for the plant, which is not possible in many real problems. Reinforcement
learning algorithms, on the other hand, are able to explore and learn to improve control performance
without knowing the plant model in advance. Therefore, applying reinforcement learning techniques
to feedback control systems could redeem the loss of performance of feedback control caused by the
imprecision of plant models.
This thesis studies how to integrate statespace models of control systems with reinforcement
|
