Convergence analysis of neural networks that solve linear programming problems

Author

Ferreira, L.V. ; Kaszkurewicz, E. ; Bhaya, A.

Author_Institution

Dept. of Electr. Eng., Univ. Fed. do Rio de Janeiro, Brazil

Volume

3

fYear

2002

fDate

6/24/1905 12:00:00 AM

Firstpage

2476

Lastpage

2481

Abstract

Artificial neural networks for solving different variants of linear programming problems are proposed and analyzed by the Lyapunov direct method. An energy function with an exact penalty term is associated with each variant and leads to a discontinuous dynamic gradient system model of an artificial neural network. The objective is to derive conditions that the network gains must satisfy in order to ensure convergence to the solution set of the linear programming problems. This objective is attained by representing the neural networks in a Persidskii-type form (S.K. Persidskii, 1969) and using an associated diagonal-type Lyapunov function

Keywords

Lyapunov methods; convergence; gradient methods; linear programming; mathematics computing; neural nets; problem solving; Lyapunov direct method; Persidskii-type representation; artificial neural networks; convergence analysis; diagonal-type Lyapunov function; discontinuous dynamic gradient system model; energy function; exact penalty term; linear programming problem solving; network gains; Analog computers; Artificial neural networks; Circuits; Constraint optimization; Convergence; Dynamic programming; Electronic mail; Linear programming; Neural networks; Optimization methods;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on

Conference_Location

Honolulu, HI

ISSN

1098-7576

Print_ISBN

0-7803-7278-6

Type

conf

DOI

10.1109/IJCNN.2002.1007531

Filename

1007531