arXiv subjects

Benjamin K Tapley

Publications and source records attributed to Benjamin K Tapley.

5 recordsLinked to original sources

arXiv2024

Symplectic Neural Networks Based on Dynamical Systems

We present and analyze a framework for designing symplectic neural networks (SympNets) based on geometric integrators for Hamiltonian differential equations. The SympNets are universal approximators in the space of Hamiltonian diffeomorphisms, interpretable and have a non-vanishing gradient property. We also give a representation theory for linear systems, meaning the proposed P-SympNets can exactly parameterize any symplectic map corresponding to quadratic Hamiltonians. Extensive numerical tests demonstrate increased expressiveness and accuracy -- often several orders of magnitude better -- for lower training cost over existing architectures. Lastly, we show how to perform symbolic Hamiltonian regression with SympNets for polynomial systems using backward error analysis.

cs.LG↗

arXiv2022

Geometric integration of ODEs using multiple quadratic auxiliary variables

We present a novel numerical method for solving ODEs while preserving polynomial first integrals. The method is based on introducing multiple quadratic auxiliary variables to reformulate the ODE as an equivalent but higher-dimensional ODE with only quadratic integrals to which the midpoint rule is applied. The quadratic auxiliary variables can subsequently be eliminated yielding a midpoint-like method on the original phase space. The resulting method is shown to be a novel discrete gradient method. Furthermore, the averaged vector field method can be obtained as a special case of the proposed method. The method can be extended to higher-order through composition and is illustrated through a number of numerical examples.

math.NA↗

arXiv2021

On the preservation of second integrals by Runge-Kutta methods

One can elucidate integrability properties of ordinary differential equations (ODEs) by knowing the existence of second integrals (also known as weak integrals or Darboux polynomials for polynomial ODEs). However, little is known about how they are preserved, if at all, under numerical methods. Here, we show that in general all Runge-Kutta methods will preserve all affine second integrals but not (irreducible) quadratic second integrals. A number of interesting corollaries are discussed, such as the preservation of certain rational integrals by arbitrary Runge-Kutta methods. We also study the special case of affine second integrals with constant cofactor and discuss the preservation of affine higher integrals.

math.NA↗

arXiv2019

Computing cost-effective particle trajectories in numerically calculated incompressible fluids using geometric methods

We present an novel algorithm for tracking massless solid particles in a divergence-free velocity field that is only available at discrete points in space and time such as those arising from a direct numerical simulation of Navier-Stokes. The algorithm creates a divergence-free approximation to the numerical field using matrix valued radial basis functions, which is integrated in time using a volume-preserving map. The resulting method is able to calculate accurate trajectories in a helical vortex using much larger step-sizes and a far lower number of interpolation points which results in a more efficient algorithm compared to a conventional scheme.

physics.comp-ph↗

arXiv2018

Symplectic integration of PDEs using Clebsch variables

Many PDEs (Burgers' equation, KdV, Camassa-Holm, Euler's fluid equations,...) can be formulated as infinite-dimensional Lie-Poisson systems. These are Hamiltonian systems on manifolds equipped with Poisson brackets. The Poisson structure is connected to conservation properties and other geometric features of solutions to the PDE and, therefore, of great interest for numerical integration. For the example of Burgers' equations and related PDEs we use Clebsch variables to lift the original system to a collective Hamiltonian system on a symplectic manifold whose structure is related to the original Lie-Poisson structure. On the collective Hamiltonian system a symplectic integrator can be applied. Our numerical examples show excellent conservation properties and indicate that the disadvantage of an increased phase-space dimension can be outweighed by the advantage of symplectic integration.

math.NA↗