In this short note, the performance of two kinds of physics-guided computing schemes, namely the Hamiltonian Neural Network and the Port-Hamiltonian Neural Network, are discussed through the predicted dynamics of two coupled Duffing oscillators. First, we propose a new error bound which holds for both types of networks. Then, we numerically investigate some alternative activation functions in terms of prediction accuracy. The numerical results show the potential of the approaches when compared to the standard neural networks in the transient regime.

Using Hamiltonian Neural Networks to Model Two Coupled Duffing Oscillators

Tomasiello S.
2023-01-01

Abstract

In this short note, the performance of two kinds of physics-guided computing schemes, namely the Hamiltonian Neural Network and the Port-Hamiltonian Neural Network, are discussed through the predicted dynamics of two coupled Duffing oscillators. First, we propose a new error bound which holds for both types of networks. Then, we numerically investigate some alternative activation functions in terms of prediction accuracy. The numerical results show the potential of the approaches when compared to the standard neural networks in the transient regime.
2023
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11386/4870511
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 1
social impact