Detail publikace

Inverse analysis based on Small-sample stochastic training of neural network

Originální název

Inverse analysis based on Small-sample stochastic training of neural network

Anglický název

Inverse analysis based on Small-sample stochastic training of neural network

Jazyk

en

Originální abstrakt

The paper suggests a new approach of inverse analysis to obtain parameters of a computational model in order to obtain best agreement witch experimental data. The proposed inverse analysis approach is based on coupling of a computational model and the stochastic training of artificial neural network. The identification parameters play the role of basic random variables with a scatter reflecting the physical range of possible values. A novelty of the approach is the utilization of efficient small-sample simulation method Latin Hypercube Sampling (LHS) used for preparation of training set utilized in stochastic training of neural network. Once the network is trained it represented an approximation consequently utilized in an opposite way: For the given experimental data to provide the best possible set of model parameters.

Anglický abstrakt

The paper suggests a new approach of inverse analysis to obtain parameters of a computational model in order to obtain best agreement witch experimental data. The proposed inverse analysis approach is based on coupling of a computational model and the stochastic training of artificial neural network. The identification parameters play the role of basic random variables with a scatter reflecting the physical range of possible values. A novelty of the approach is the utilization of efficient small-sample simulation method Latin Hypercube Sampling (LHS) used for preparation of training set utilized in stochastic training of neural network. Once the network is trained it represented an approximation consequently utilized in an opposite way: For the given experimental data to provide the best possible set of model parameters.

BibTex


@inproceedings{BUT21414,
  author="Drahomír {Novák} and David {Lehký}",
  title="Inverse analysis based on Small-sample stochastic training of neural network",
  annote="The paper suggests a new approach of inverse analysis to obtain parameters of a computational model in order to obtain best agreement witch experimental data. The proposed inverse analysis approach is based on coupling of a computational model and the stochastic training of artificial neural network. The identification parameters play the role of basic random variables with a scatter reflecting the physical range of possible values. A novelty of the approach is the utilization of efficient small-sample simulation method Latin Hypercube Sampling (LHS) used for preparation of training set utilized in stochastic training of neural network. Once the network is trained it represented an approximation consequently utilized in an opposite way: For the given experimental data to provide the best possible set of model parameters.",
  address="Stéphane Lecoeuche and Dimitris Tsaptsinos",
  booktitle="Novel Applications of Neural Network in Engineering",
  chapter="21414",
  institution="Stéphane Lecoeuche and Dimitris Tsaptsinos",
  year="2005",
  month="august",
  pages="155--162",
  publisher="Stéphane Lecoeuche and Dimitris Tsaptsinos",
  type="conference paper"
}