Lagrange programming neural networks

Author

Zhang, Shengwei ; Constantinides, A.G.

Author_Institution

Expervision, San Jose, CA, USA

Volume

39

Issue

7

fYear

1992

fDate

7/1/1992 12:00:00 AM

Firstpage

441

Lastpage

452

Abstract

A class of neural networks appropriate for general nonlinear programming, i.e., problems including both equality and inequality constraints, is analyzed in detail. The methodology is based on the Lagrange multiplier theory in optimization and seeks to provide solutions satisfying the necessary conditions of optimality. The equilibrium point of the network satisfies the Kuhn-Tucker condition for the problem. No explicit restriction is imposed on the form of the cost function apart from some general regularity and convexity conditions. The stability of the neural networks is analyzed in detail. The transient behavior of the network is simulated and the validity of the approach is verified for a practical problem, maximum entropy image restoration

Keywords

image reconstruction; neural nets; nonlinear programming; stability; Kuhn-Tucker condition; Lagrange multiplier theory; Lagrange programming; convexity; cost function; equality constraints; equilibrium point; inequality constraints; maximum entropy image restoration; neural networks; nonlinear programming; optimization; regularity; stability; transient behavior; Analog computers; Circuits; Computational modeling; Computer networks; Constraint optimization; Cost function; Lagrangian functions; Neural networks; Optimization methods; Stability analysis;

fLanguage

English

Journal_Title

Circuits and Systems II: Analog and Digital Signal Processing, IEEE Transactions on

Publisher

ieee

ISSN

1057-7130

Type

jour

DOI

10.1109/82.160169

Filename

160169