On the numerical analysis and visualisation of implicit ordinary differential equations (Q782675)

From MaRDI portal
Revision as of 12:04, 30 January 2024 by Import240129110113 (talk | contribs) (Added link to MaRDI item.)
scientific article
Language Label Description Also known as
English
On the numerical analysis and visualisation of implicit ordinary differential equations
scientific article

    Statements

    On the numerical analysis and visualisation of implicit ordinary differential equations (English)
    0 references
    0 references
    0 references
    0 references
    28 July 2020
    0 references
    Much work has been published on the numerical integration of implicit ODEs and DAEs from the eighties, by Gear, Rabier and Rheinboldt or Tuomela. In this paper, the differential equation of order \(q>0\) reads \[F(x,u, \dot{u},\dots,u^{(q)})=0,\] The authors apply results of the geometrical approach to compute the solutions near a singular point, that a value \((x_0,u_0)\in \mathbb{R}^{q+1}\) such that \((\partial F/\partial u^q)(x_0,u_0)=0\). Then, an original method is developed to give a comprehensive visualization of the phase portrait in dimension \(2(.5)\) and \(3\). More precisely, the authors consider a differential equation as a submanifold \(\mathcal{R}_q\) of the jet bundle \(J_q(\pi)\) of order \(q\), at a point \(\rho\) and use the Vessiot distribution at this point, i.e. the part of the contact structure that is tangent to \(\mathcal{R}_q\). Singular points can be classified with V.I. Arnol'd's terminology. Regular ones are those having a Vessiot distribution of dimension \(1\). Its is known that the Vessiot distribution is locally equivalent to a vector field which at standard points or regular singular ones, features one general solution (computed as the graph of a function), while the irregular singularities become stationary points of this vector field. For the numerical integration, the standard points and the regular singularities can be tackled by numerical integration of the Vessiot distribution. The geometrical approach and the introduction of general solutions show that at irregular singularities, the numerical computation of the invariant manifold \(W\) would lead to the phase portrait. This is a difficult problem in higher dimension. The authors manage this case by a combination of symbolic computation and numerical integration, implementing the algorithms of \textit{W.-J. Beyn} and \textit{W. Kleß} [Numer. Math. 80, No. 1, 1--38 (1998; Zbl 0909.65044)] and \textit{T. Eirola} and \textit{J. von Pfaler} [Numer. Math. 99, No. 1, 25--46 (2004; Zbl 1063.65139)] in the Matlab system. This is illustrated by non trivial numerical examples, including the visualisation. The paper is well-structured and easy to read, really interesting as it makes more concrete the idea we may have of the complicated phase portrait of an implicit DE near an impasse point or a non trivial singularity.
    0 references
    implicit differential equations
    0 references
    singularities
    0 references
    Vessiot distribution
    0 references
    numerical integration
    0 references
    visualisation
    0 references
    impasse points
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references