On second order nonlinear boundary value problems and the distributional Henstock-Kurzweil integral (Q2345978)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On second order nonlinear boundary value problems and the distributional Henstock-Kurzweil integral
scientific article

    Statements

    On second order nonlinear boundary value problems and the distributional Henstock-Kurzweil integral (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    28 May 2015
    0 references
    The paper is concerned with the second-order boundary value problem \[ -x''(t)=f(t,x(t)),\quad t\in[0,1],\eqno(1) \] \[ x'(0)=0,\quad \beta x'(1)+x(\eta)=0,\eqno(2) \] where \(\beta>0\) and \(\eta\in[0,1]\). Instead of dealing with classical solutions, the authors consider the case when \(x\) is merely continuous, \(x'\) is a continuous function which is the distributional derivative of \(x\), and \(x''\) is a distribution which is the distributional derivative of \(x'\). The right-hand side \(f\) is also a distribution. It is somewhat misleading that the distributional equation is still written in the form (1), since the symbol \(x''(t)\) is meaningless when \(x''\) is a distribution. According to the reviewer's opinion, the correct notation should be \(-x''=f(x)\), where \(f\) is a mapping from \(\mathcal{C}([0,1])\) to the space of distributions on \([0,1]\). The authors assume that the following conditions are satisfied: {\parindent=6mm \begin{itemize}\item[1)] For each \(x\in \mathcal{C}([0,1])\), \(f(x)\) is distributionally Henstock-Kurzweil integrable in the sense of Erik Talvila's paper ``The distributional Denjoy integral'' [\textit{E. Talvila}, Real Anal. Exch. 33, No. 1, 51--82 (2008; Zbl 1154.26011)]. \item[2)] \(f\) is continuous in \(x\). \item[3)] There exist distributions \(f_-\), \(f_+\) on \([0,1]\) which are distributionally Henstock-Kurzweil integrable, and such that \(\int_0^t f_-\leq \int_0^t f(x)\leq \int_0^t f_+\) for each \(x\in \mathcal{C}([0,1])\) and \(t\in[0,1]\). \end{itemize}} Using Schauder's fixed point theorem, it is proved that the distributional boundary value problem has a~solution. The result is illustrated on a simple example.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    nonlinear boundary value problem
    0 references
    distributional derivative
    0 references
    Schauder's fixed point theorem
    0 references
    distributional Henstock-Kurzweil integral
    0 references
    0 references
    0 references