quotation:		[Copy]
		James Nate KNIGHT,Charles ANDERSON.[en_title][J].Control Theory and Technology,2011,9(3):410~420.[Copy]

This Paper:Browse 4249 Download 707	码上扫一扫！

JamesNateKNIGHT,CharlesANDERSON
0 Fontlarge +\|Default\|Small
(Department of Computer Science, Colorado State University)

摘要:

关键词:

DOI：

Received:July 28, 2010Revised:March 22, 2011

基金项目:This work was supported by the National Natural Science Foundation (No.0245291).

Stable reinforcement learning with recurrent neural networks

James Nate KNIGHT,Charles ANDERSON

(Department of Computer Science, Colorado State University)

Abstract:

In this paper, we present a technique for ensuring the stability of a large class of adaptively controlled systems. We combine IQC models of both the controlled system and the controller with a method of filtering control parameter updates to ensure stable behavior of the controlled system under adaptation of the controller. We present a specific application to a system that uses recurrent neural networks adapted via reinforcement learning techniques. The work presented extends earlier works on stable reinforcement learning with neural networks. Specifically, we apply an improved IQC analysis for RNNs with time-varying weights and evaluate the approach on more complex control system.

Key words: Stability analysis Integral quadratic constraint Recurrent neural network Reinforcement learning Linear matrix inequality