Nonlinear maps of convex sets in Hilbert spaces with application to kinetic equations (Q734072)

From MaRDI portal
Revision as of 01:23, 2 July 2024 by ReferenceBot (talk | contribs) (‎Changed an Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Nonlinear maps of convex sets in Hilbert spaces with application to kinetic equations
scientific article

    Statements

    Nonlinear maps of convex sets in Hilbert spaces with application to kinetic equations (English)
    0 references
    0 references
    19 October 2009
    0 references
    Let \(\mathcal H\) be a separable Hilbert space, \(\mathcal U\subset \mathcal H\) open convex and \(f:\mathcal U\to \mathcal H\) a smooth map. Let also \(\Omega\) be an open convex subset of \(\mathcal H\) such that \(\overline{\Omega}\subset \mathcal U\). The paper is concerned with the following two problems: {\parindent5mm \begin{itemize}\item[1.] Supposing \(f\) Lipschitz, find sufficient conditions such that for sufficiently small \(\varepsilon\) (depending on Lip\((f)\)), the image \((I+\varepsilon f)(\Omega)\) of \(\Omega\) by \(I+\varepsilon f\) is convex. \item[2.] Supposing \(df(u):\mathcal H\to\mathcal H\) symmetrizable with the spectrum \(\sigma(df(u))\) contained in \((0,\infty)\), find sufficient conditions in order that \(f(\Omega)\) is convex. \end{itemize}} It turns out that, as in the finite-dimensional case, the most important condition relating \(f\) and \(\partial \Omega\) is that the tangent space \(T_\omega(\partial \Omega)\) to \(\partial \Omega\) at \(\omega\) satisfies \(df(\omega)(T_\omega(\partial \Omega)) \subset T_\omega(\partial \Omega)\) for every \(\omega\in\partial\Omega\). The main tool used to overcome the difficulties specific to the infinite-dimensional case is the concept of standard Fredholm operator, meaning an operator of the form \(c I+K\) with \(c\geq 0\) and \(K\) compact. One supposes that \(df(\omega)\) and \(d^2f(\omega)\) are standard Fredholm operators on \(\mathcal H_0=T_\omega(\partial \Omega)\) for every \(w\in\partial \Omega\). As applications, the author shows that the first result gives an invariance principle for finite difference schemes for nonlinear differential equations in Hilbert spaces, while the second one provides an invariance principle for certain convex sets in an \(L^2\)-space under the flow of a class of kinetic transport equations. The obtained results extend to the infinite-dimensional setting those obtained by the author in the finite-dimensional case [cf. Arch. Ration. Mech. Anal. 160, No.~3, 245--269 (2001; Zbl 0993.65096) and Arch. Ration. Mech. Anal. 171, No.~2, 297--299 (2004; Zbl 1059.65073)].
    0 references
    convex sets
    0 references
    invariant domains
    0 references
    Hilbert space
    0 references
    Lipschitz maps
    0 references
    smooth maps
    0 references
    finite difference schemes
    0 references
    kinetic equations
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references