Skip to Main Content (Press Enter)

Logo CNR
  • ×
  • Home
  • Persone
  • Pubblicazioni
  • Strutture
  • Competenze

UNI-FIND
Logo CNR

|

UNI-FIND

cnr.it
  • ×
  • Home
  • Persone
  • Pubblicazioni
  • Strutture
  • Competenze
  1. Pubblicazioni

Spectrum-Free and Meshless Solvers of Parabolic PDEs

Articolo
Data di Pubblicazione:
2021
Abstract:
We propose a novel collocation method with Radial Basis Functions for the solution of the inhomogeneous parabolic equation (?t + L)u(·, t) = f on ? Rd , with L elliptic operator. As original contribution, we rewrite the solution in terms of the exponential operator exp(-tL), which is then computed through the Padè-Chebyshev approximation of the 1D Gaussian function. The resulting meshless solver uniformly converges to the ground-truth solution, as the degree of the rational polynomial increases, and is independent of the evaluation of the spectrumof L (i.e., spectrum-free), of the discretisation of the temporal derivative, and of user-defined parameters. Since the solution is approximated as a linear combination of Radial Basis Functions, we study the conditions on the generating kernel that guarantee the L-differentiability of the meshless solution. In our tests, we compare the proposed meshless and spectrum-free solvers with the meshless spectral eigen-decomposition and the meshless ?-method on the heat equation in a transient regime. With respect to these previous works, at small scales the Padè-Chebyshev method has a higher numerical stability and approximation accuracy, which are expressed in terms of the selected degree of the rational polynomial and of the spectral properties of the matrix that discretises the parabolic operator.
Tipologia CRIS:
01.01 Articolo in rivista
Keywords:
Parabolic PDEs; Heat equation; Collocation methods; Radial basis f; Spectrum; Spectrum-free solvers
Elenco autori:
Patane', Giuseppe
Autori di Ateneo:
PATANE' GIUSEPPE
Link alla scheda completa:
https://iris.cnr.it/handle/20.500.14243/441979
Pubblicato in:
JOURNAL OF SCIENTIFIC COMPUTING
Journal
  • Dati Generali

Dati Generali

URL

https://link.springer.com/article/10.1007/s10915-021-01604-x
  • Utilizzo dei cookie

Realizzato con VIVO | Designed by Cineca | 26.5.0.0 | Sorgente dati: PREPROD (Ribaltamento disabilitato)