The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Lors de la résolution de problèmes d'optimisation combinatoire avec un réseau neuronal binaire de type Hopfield, le processus de mise à jour dans le réseau neuronal est une étape importante dans la recherche d'une solution. Dans cette lettre, nous proposons une nouvelle procédure de mise à jour dans un réseau neuronal binaire de type Hopfield pour résoudre efficacement des problèmes d'optimisation combinatoire. Dans la nouvelle procédure de mise à jour, une fois que le neurone est dans un état excitateur, alors son potentiel d'entrée est en saturation positive où le potentiel d'entrée ne peut qu'être réduit mais ne peut pas être augmenté, et une fois que le neurone est dans un état inhibiteur, alors son potentiel d'entrée est en saturation négative où le potentiel d'entrée ne peut qu'être augmenté mais ne peut pas être réduit. La nouvelle procédure de mise à jour est évaluée et comparée à la procédure originale et à d'autres méthodes améliorées grâce à des simulations basées sur le problème N-Queens. Les résultats montrent que la nouvelle procédure de mise à jour améliore la capacité de recherche des réseaux de neurones avec un temps de calcul plus court. En particulier, les résultats de la simulation montrent que les performances de la méthode proposée surpassent celles des méthodes existantes pour le problème à N-reines dans un modèle de calcul parallèle synchrone.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Rong-Long WANG, Zheng TANG, Qi-Ping CAO, "A New Updating Procedure in the Hopfield-Type Network and Its Application to N-Queens Problem" in IEICE TRANSACTIONS on Fundamentals,
vol. E85-A, no. 10, pp. 2368-2372, October 2002, doi: .
Abstract: When solving combinatorial optimization problems with a binary Hopfield-type neural network, the updating process in neural network is an important step in achieving a solution. In this letter, we propose a new updating procedure in binary Hopfield-type neural network for efficiently solving combinatorial optimization problems. In the new updating procedure, once the neuron is in excitatory state, then its input potential is in positive saturation where the input potential can only be reduced but cannot be increased, and once the neuron is in inhibitory state, then its input potential is in negative saturation where the input potential can only be increased but cannot be reduced. The new updating procedure is evaluated and compared with the original procedure and other improved methods through simulations based on N-Queens problem. The results show that the new updating procedure improves the searching capability of neural networks with shorter computation time. Particularly, the simulation results show that the performance of proposed method surpasses the exiting methods for N-queens problem in synchronous parallel computation model.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e85-a_10_2368/_p
Copier
@ARTICLE{e85-a_10_2368,
author={Rong-Long WANG, Zheng TANG, Qi-Ping CAO, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={A New Updating Procedure in the Hopfield-Type Network and Its Application to N-Queens Problem},
year={2002},
volume={E85-A},
number={10},
pages={2368-2372},
abstract={When solving combinatorial optimization problems with a binary Hopfield-type neural network, the updating process in neural network is an important step in achieving a solution. In this letter, we propose a new updating procedure in binary Hopfield-type neural network for efficiently solving combinatorial optimization problems. In the new updating procedure, once the neuron is in excitatory state, then its input potential is in positive saturation where the input potential can only be reduced but cannot be increased, and once the neuron is in inhibitory state, then its input potential is in negative saturation where the input potential can only be increased but cannot be reduced. The new updating procedure is evaluated and compared with the original procedure and other improved methods through simulations based on N-Queens problem. The results show that the new updating procedure improves the searching capability of neural networks with shorter computation time. Particularly, the simulation results show that the performance of proposed method surpasses the exiting methods for N-queens problem in synchronous parallel computation model.},
keywords={},
doi={},
ISSN={},
month={October},}
Copier
TY - JOUR
TI - A New Updating Procedure in the Hopfield-Type Network and Its Application to N-Queens Problem
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 2368
EP - 2372
AU - Rong-Long WANG
AU - Zheng TANG
AU - Qi-Ping CAO
PY - 2002
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E85-A
IS - 10
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - October 2002
AB - When solving combinatorial optimization problems with a binary Hopfield-type neural network, the updating process in neural network is an important step in achieving a solution. In this letter, we propose a new updating procedure in binary Hopfield-type neural network for efficiently solving combinatorial optimization problems. In the new updating procedure, once the neuron is in excitatory state, then its input potential is in positive saturation where the input potential can only be reduced but cannot be increased, and once the neuron is in inhibitory state, then its input potential is in negative saturation where the input potential can only be increased but cannot be reduced. The new updating procedure is evaluated and compared with the original procedure and other improved methods through simulations based on N-Queens problem. The results show that the new updating procedure improves the searching capability of neural networks with shorter computation time. Particularly, the simulation results show that the performance of proposed method surpasses the exiting methods for N-queens problem in synchronous parallel computation model.
ER -