Communication Dans Un Congrès Année : 2020

Dynamical System Identification by Bayesian Inference

Robert K. Niven
  • Fonction : Auteur
  • PersonId : 1086885
Mohammad-Djafari Ali
Markus Abel
  • Fonction : Auteur
  • PersonId : 1086886
Markus Quade
  • Fonction : Auteur

Résumé

Fluid flow systems provide physical expressions of dynamical systems, typically writtenẋ x x = f (x x x), where x x x is the state vector and f is the system function or model. The identification of the dynamical system f from time-series data {x 1 , ..., x n }, an inverse problem, is a long-held challenge. Historically, this has been examined by linear or nonlinear regression, convolution methods, neural networks or evolutionary computation, but these mostly lie outside the rigorous framework of Bayesian inference. Here we examine the maximum a-posteriori (MAP) Bayesian method for system identification, which is shown to be equivalent to Tikhonov regularization, and in fact provides sound theoretical justifications for the choices of residual and regularization terms. The joint maximum a-posteriori (JMAP) and variational Bayesian approximation (VBA) are demonstrated by comparison to the popular SINDy regularization method, by application to the Rössler dynamical system.

Fichier principal
Vignette du fichier
afmc2020-latex-Niven_etal_paper244_Bayes_v3.pdf (1.35 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-03088615 , version 1 (26-12-2020)

Licence

Identifiants

Citer

Robert K. Niven, Mohammad-Djafari Ali, Laurent Cordier, Markus Abel, Markus Quade. Dynamical System Identification by Bayesian Inference. 22nd Australasian Fluid Mechanics Conference AFMC2020, Dec 2020, Brisbane, Australia. ⟨10.14264/692fcb8⟩. ⟨hal-03088615⟩
306 Consultations
447 Téléchargements

Altmetric

Partager

  • More