Gradient-based Neural Network for Online Solution of Lyapunov Matrix Equation with Li Activation Function

DOI

10.2991/icitmi-15.2015.161How to use a DOI?

Keywords

Gradient-based neural network, Laypunov matrix equation, Activation function

Abstract

A new type of activation function, named Li activation function, is used in gradient-based neural network (GNN) to solve Lyapunov matrix equation. With this activation function, theoretical analysis shows that GNN can converge in finite time, while it can converge only in infinite time with two conventional activation functions — linear and power-sigmoid. Computer simulation results confirm that GNN with Li activation function can not only globally converge to the solution of the Lyapunov matrix equation but also converge in finite time. GNN with the conventional two activation functions are also simulated as a contrast.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Shiheng Wang
AU  - Shidong Dai
AU  - Ke Wang
PY  - 2015/10
DA  - 2015/10
TI  - Gradient-based Neural Network for Online Solution of Lyapunov Matrix Equation with Li Activation Function
BT  - Proceedings of the 4th International Conference on Information Technology and Management Innovation
PB  - Atlantis Press
SP  - 955
EP  - 959
SN  - 2352-538X
UR  - https://doi.org/10.2991/icitmi-15.2015.161
DO  - 10.2991/icitmi-15.2015.161
ID  - Wang2015/10
ER  -