A decision-making machine learning approach in Hermite spectral approximations of partial differential equations

DOI10.1007/S10915-022-01853-4zbMATH Open1490.65297arXiv2104.02181OpenAlexW3147774872WikidataQ114225565 ScholiaQ114225565MaRDI QIDQ2149019FDOQ2149019

Daniele Funaro, Lorella Fatone, Gianmarco Manzini

Publication date: 24 June 2022

Published in: Journal of Scientific Computing (Search for Journal in Brave)

Abstract: The accuracy and effectiveness of Hermite spectral methods for the numerical discretization of partial differential equations on unbounded domains, are strongly affected by the amplitude of the Gaussian weight function employed to describe the approximation space. This is particularly true if the problem is under-resolved, i.e., there are no enough degrees of freedom. The issue becomes even more crucial when the equation under study is time-dependent, forcing in this way the choice of Hermite functions where the corresponding weight depends on time. In order to adapt dynamically the approximation space, it is here proposed an automatic decision-making process that relies on machine learning techniques, such as deep neural networks and support vector machines. The algorithm is numerically tested with success on a simple 1D problem, but the main goal is its exportability in the context of more serious applications.

Full work available at URL: https://arxiv.org/abs/2104.02181

zbMATH Keywords

machine learningsupport vector machineneural networksspectral methodsgeneralized Hermite functionstime-dependent heat equation

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Vlasov equations (35Q83) Spectral, collocation and related methods for boundary value problems involving PDEs (65N35)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• DGM: a deep learning algorithm for solving partial differential equations
• Kernel methods in machine learning
• Time-dependent Hermite-Galerkin spectral method and its applications
• Spectral Methods
• On the kinetic theory of rarefied gases
• Approximation by superpositions of a sigmoidal function
• Estimation of dependences based on empirical data. Transl. from the Russian by Samuel Kotz
• The Hermite Spectral Method for Gaussian-Type Functions
• Vlasov simulations using velocity-scaled Hermite representations
• Arbitrary-order time-accurate semi-Lagrangian spectral approximations of the Vlasov-Poisson system
• Approximation of Some Diffusion Evolution Equations in Unbounded Domains by Hermite Functions
• Error estimation of Hermite spectral method for nonlinear partial differential equations
• Hermite Spectral Methods with a Time-Dependent Scaling for Parabolic Equations in Unbounded Domains
• Mixed Spectral Method for Heat Transfer Using Generalised Hermite Functions and Legendre Polynomials
• Approximation of functions and their derivatives: A neural network implementation with applications
• Stability and conservation properties of Hermite-based approximations of the Vlasov-Poisson system
• On the use of Hermite functions for the Vlasov-Poisson system
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• An introduction to neural network methods for differential equations
• Convergence of Spectral Discretizations of the Vlasov--Poisson System
• Multi-dimensional, fully-implicit, spectral method for the Vlasov-Maxwell equations with exact conservation laws in discrete form
• A Legendre-Fourier spectral method with exact conservation laws for the Vlasov-Poisson system
• A semi-Lagrangian spectral method for the Vlasov-Poisson system based on Fourier, Legendre and Hermite polynomials
• nPINNs: nonlocal physics-informed neural networks for a parametrized nonlocal universal Laplacian operator. Algorithms and applications
• On the Convergence of Physics Informed Neural Networks for Linear Second-Order Elliptic and Parabolic Type PDEs
• Numerical Simulation of Microflows Using Hermite Spectral Methods

Cited In (2)

• Anti-symmetric and positivity preserving formulation of a spectral method for Vlasov-Poisson equations
• Physics-based adaptivity of a spectral method for the Vlasov-Poisson equations based on the asymmetrically-weighted Hermite expansion in velocity space

Uses Software

• KELLEY
• LIBSVM
• SpectralPlasmaSolver
• DGM
• NSFnets
• hp-VPINNs

This page was built for publication: A decision-making machine learning approach in Hermite spectral approximations of partial differential equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2149019)