Presentation 1996/2/3
Theoretical analysis and simulations of Hopfield neural networks using Lagrange multipliers
Kenji MURAKAMI, Takahumi OOHORI, Kazuhisa WATANABE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The augmented penalty function using the Lagrange multipliers was proposed to improve the feasibility of the Hopfield network with the penalty function. However, it was pointed that the convergence to the feasible solutions is not assured when the penalty factor r is small. Computational results for several constrained combinatorial problems show that, for the problems with the linear objective, any r&ge0 can make the local minima feasible by updating Lagrange multipliers and, for the problems with the quadratic objective, there exist the r's that can make the local minima alway feasibe or alway infeasible by updating Lagrange multipliers. We also theoretically analyze the above results for the WTA(winner-takes-all)problems.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) interconnected neural network / combinatorial optimization problem / Lagrange multiplier / augmented penalty function / theoretical analysis / winner-takes-all problem
Paper # NC95-110
Date of Issue

Conference Information
Committee NC
Conference Date 1996/2/3(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Theoretical analysis and simulations of Hopfield neural networks using Lagrange multipliers
Sub Title (in English)
Keyword(1) interconnected neural network
Keyword(2) combinatorial optimization problem
Keyword(3) Lagrange multiplier
Keyword(4) augmented penalty function
Keyword(5) theoretical analysis
Keyword(6) winner-takes-all problem
1st Author's Name Kenji MURAKAMI
1st Author's Affiliation Department of Electrical Engineering, Hokkaido Institute of Technology()
2nd Author's Name Takahumi OOHORI
2nd Author's Affiliation Department of Electrical Engineering, Hokkaido Institute of Technology
3rd Author's Name Kazuhisa WATANABE
3rd Author's Affiliation Department of Electrical Engineering, Hokkaido Institute of Technology
Date 1996/2/3
Paper # NC95-110
Volume (vol) vol.95
Number (no) 506
Page pp.pp.-
#Pages 8
Date of Issue