Hessian matrix estimation in hybrid systems based on an embedded FFNN

Seung Mook Baek; Jung Wook Park

doi:10.1109/TNN.2010.2042728

Hessian matrix estimation in hybrid systems based on an embedded FFNN

Seung Mook Baek, Jung Wook Park

Department of Electrical and Electronic Engineering

Research output: Contribution to journal › Article › peer-review

5 Citations (Scopus)

Abstract

This paper describes the Hessian matrix estimation of nonsmooth nonlinear parameters by the identifier based on a feedforward neural network (FFNN) embedded in a hybrid system, which is modeled by the differentialalgebraicimpulsiveswitched (DAIS) structure. After identifying full dynamics of the hybrid system, the FFNN is used to estimate second-order derivatives of an objective function J with respect to the nonlinear parameters from the gradient information, which are trajectory sensitivities. Then, the estimated Hessian matrix is applied to the optimal tuning of a saturation limiter used in a practical engineering system.

Original language	English
Article number	5431074
Pages (from-to)	1533-1542
Number of pages	10
Journal	IEEE Transactions on Neural Networks
Volume	21
Issue number	10
DOIs	https://doi.org/10.1109/TNN.2010.2042728
Publication status	Published - 2010 Oct

Bibliographical note

Funding Information:
Manuscript received June 02, 2007; revised June 22, 2009 and December 18, 2009; accepted December 28, 2009. Date of publication March 15, 2010; date of current version September 01, 2010. This work was supported by the Manpower Development Program for Energy & Resources of MKE with Yonsei Electric Power Research Center (YEPRC) at Yonsei University, Seoul, Korea.

All Science Journal Classification (ASJC) codes

Software
Computer Science Applications
Computer Networks and Communications
Artificial Intelligence

Access to Document

10.1109/TNN.2010.2042728

Cite this

@article{2e60be0126b54e69933c1aacab265862,

title = "Hessian matrix estimation in hybrid systems based on an embedded FFNN",

abstract = "This paper describes the Hessian matrix estimation of nonsmooth nonlinear parameters by the identifier based on a feedforward neural network (FFNN) embedded in a hybrid system, which is modeled by the differentialalgebraicimpulsiveswitched (DAIS) structure. After identifying full dynamics of the hybrid system, the FFNN is used to estimate second-order derivatives of an objective function J with respect to the nonlinear parameters from the gradient information, which are trajectory sensitivities. Then, the estimated Hessian matrix is applied to the optimal tuning of a saturation limiter used in a practical engineering system.",

author = "Baek, {Seung Mook} and Park, {Jung Wook}",

note = "Funding Information: Manuscript received June 02, 2007; revised June 22, 2009 and December 18, 2009; accepted December 28, 2009. Date of publication March 15, 2010; date of current version September 01, 2010. This work was supported by the Manpower Development Program for Energy & Resources of MKE with Yonsei Electric Power Research Center (YEPRC) at Yonsei University, Seoul, Korea.",

year = "2010",

month = oct,

doi = "10.1109/TNN.2010.2042728",

language = "English",

volume = "21",

pages = "1533--1542",

journal = "IEEE Transactions on Neural Networks",

issn = "1045-9227",

publisher = "IEEE Computational Intelligence Society",

number = "10",

}

TY - JOUR

T1 - Hessian matrix estimation in hybrid systems based on an embedded FFNN

AU - Baek, Seung Mook

AU - Park, Jung Wook

N1 - Funding Information: Manuscript received June 02, 2007; revised June 22, 2009 and December 18, 2009; accepted December 28, 2009. Date of publication March 15, 2010; date of current version September 01, 2010. This work was supported by the Manpower Development Program for Energy & Resources of MKE with Yonsei Electric Power Research Center (YEPRC) at Yonsei University, Seoul, Korea.

PY - 2010/10

Y1 - 2010/10

N2 - This paper describes the Hessian matrix estimation of nonsmooth nonlinear parameters by the identifier based on a feedforward neural network (FFNN) embedded in a hybrid system, which is modeled by the differentialalgebraicimpulsiveswitched (DAIS) structure. After identifying full dynamics of the hybrid system, the FFNN is used to estimate second-order derivatives of an objective function J with respect to the nonlinear parameters from the gradient information, which are trajectory sensitivities. Then, the estimated Hessian matrix is applied to the optimal tuning of a saturation limiter used in a practical engineering system.

AB - This paper describes the Hessian matrix estimation of nonsmooth nonlinear parameters by the identifier based on a feedforward neural network (FFNN) embedded in a hybrid system, which is modeled by the differentialalgebraicimpulsiveswitched (DAIS) structure. After identifying full dynamics of the hybrid system, the FFNN is used to estimate second-order derivatives of an objective function J with respect to the nonlinear parameters from the gradient information, which are trajectory sensitivities. Then, the estimated Hessian matrix is applied to the optimal tuning of a saturation limiter used in a practical engineering system.

UR - http://www.scopus.com/inward/record.url?scp=77957772394&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77957772394&partnerID=8YFLogxK

U2 - 10.1109/TNN.2010.2042728

DO - 10.1109/TNN.2010.2042728

M3 - Article

C2 - 20236883

AN - SCOPUS:77957772394

SN - 1045-9227

VL - 21

SP - 1533

EP - 1542

JO - IEEE Transactions on Neural Networks

JF - IEEE Transactions on Neural Networks

IS - 10

M1 - 5431074

ER -

Hessian matrix estimation in hybrid systems based on an embedded FFNN

Abstract

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this