Dissipative Learning Method for Delayed Neural Networks
- 주제(키워드) Neural networks
- 발행기관 고려대학교 대학원
- 지도교수 안춘기
- 발행년도 2016
- 학위수여년월 2016. 8
- 학위구분 석사
- 학과 대학원 전기전자공학과
- 원문페이지 60 p
- 실제URI http://www.dcollection.net/handler/korea/000000068897
- 본문언어 영어
- 제출원본 000045881605
초록/요약
This thesis is concerned with the problem of (Q, S,R)-α-dissipative learning for continuous-time delayed neural networks. A new learning method is established to guarantee the asymptotic stability as well as the (Q, S,R)-α-dissipativity of the system. The result encompasses some special cases, such as H∞ performance, passivity performance, and mixed H∞/passivity performance in a unified framework. Based on the Bessel-Legendre inequality, a new set of delay-dependent sufficient linear matrix inequalities (LMIs) conditions and learning laws are presented. Numerical examples are given to illustrate the effectiveness of the proposed learning method.
more목차
1 Introduction 1
1.1 Research Background . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Outline of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Preliminaries 6
2.1 Dissipativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Bessel-Legendre Inequality . . . . . . . . . . . . . . . . . . . . . . 9
3 Dissipative Stable Learning Method for Delayed Neural Net-
works 12
3.1 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 (Q, S,R)-α-Dissipativity . . . . . . . . . . . . . . . . . . . . . . . 13
3.3 Learning Law Derivation with Bessel-Lengendre Inequality . . . 14
4 Numerical Examples 33
4.1 Example 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.2 Example 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5 Conclusion 45
6 Bibliography 47
