1998
DOI: 10.1016/s0266-352x(97)00034-7
|View full text |Cite
|
Sign up to set email alerts
|

New nested adaptive neural networks (NANN) for constitutive modeling

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
115
0
1

Year Published

2002
2002
2024
2024

Publication Types

Select...
9
1

Relationship

0
10

Authors

Journals

citations
Cited by 227 publications
(116 citation statements)
references
References 6 publications
0
115
0
1
Order By: Relevance
“…We mention for example the so called self-learning finite element procedure introduced first by Ghaboussi in [4][5][6] and developed then by Shin and Pande in [7,8] as a very promising technique of data acquisition for the ANN training process. Presentation of this method that was applied immediately to problems of diaphragma wall dimensioning would exceed the Some aspects of application of artificial neural network for numerical modeling in civil engineering frame of this paper.…”
Section: 2mentioning
confidence: 99%
“…We mention for example the so called self-learning finite element procedure introduced first by Ghaboussi in [4][5][6] and developed then by Shin and Pande in [7,8] as a very promising technique of data acquisition for the ANN training process. Presentation of this method that was applied immediately to problems of diaphragma wall dimensioning would exceed the Some aspects of application of artificial neural network for numerical modeling in civil engineering frame of this paper.…”
Section: 2mentioning
confidence: 99%
“…At worst, the data provided to train the network are biased. However, ANNs can always be updated to obtain better results by presenting additional training examples as new data become available (Ghaboussi & Sidarta 1998).…”
Section: Context and Literature Reviewmentioning
confidence: 99%
“…Usually, the number of neurons in hidden layers remains constant for a given training trial and is modified as necessary if the performance of the network is not satisfactory. However, the search of an optimal topology can be automatized using a dynamic update of the architecture during the training process [4,33]. In our approach, optimal network topologies have been found with a few trial training runs.…”
Section: Nn Trainingmentioning
confidence: 99%