Conditions for the qualified convergence of finite difference methods and the quasi-reversibility method for solving linear ill-posed Cauchy problems in a Hilbert space (Q2191539)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Conditions for the qualified convergence of finite difference methods and the quasi-reversibility method for solving linear ill-posed Cauchy problems in a Hilbert space
scientific article

    Statements

    Conditions for the qualified convergence of finite difference methods and the quasi-reversibility method for solving linear ill-posed Cauchy problems in a Hilbert space (English)
    0 references
    25 June 2020
    0 references
    This paper is aimed at studying finite difference methods and the quasi-reversibility method for the abstract Cauchy problem \[ \frac{d x(t)}{dt}= \mathcal{A}\, x(t),\quad x(0)=f, \] where \(\mathcal{A}\) is an unbounded self-adjoint linear operator mapping in a Hilbert space \(H\) with spectrum \(\sigma(\mathcal{A})\) belonging to the interval \([a,+\infty)\) for \(a>0\). In this context, this is in some sense a continuation of two papers exploiting the same approach, published by the author in 2013 and 2015 in Russian journals [\textit{M. M. Kokurin}, ``Optimization of convergence rate estimates for some classes of finite difference schemes for solving Ill-posed Cauchy problems'', Vychisl. Metody Program. 14, No. 1, 58--76 (2013), \url{https://en.num-meth.ru/index.php/journal/article/view/583}; Comput. Math. Math. Phys. 55, No. 12, 1986--2000 (2015; Zbl 1339.65071); translation from Zh. Vychisl. Mat. Mat. Fiz. 55, No. 12, 2027--2041 (2015)]. Section 2 presents `direct theorems' with respect to the convergence of finite difference methods, whilst Sections 3 and 4 deal with the corresponding `inverse theorems', which include converse results and discuss source conditions and saturation phenomena. The classical quasi-reversibility method developed by \textit{R. Lattès} and \textit{J.-L. Lions} [Méthode de quasi-réversibilité et applications. Moskva: Izdatel'stvo ``Mir'' (1970; Zbl 0197.13803)] and based on the surrogate problem \[ \frac{d x_\varepsilon(t)}{dt}= (\mathcal{A}-\varepsilon\mathcal{A}^2)\, x_\varepsilon(t),\quad x_\epsilon(0)=f, \] is considered in Section 5. For this approach, also convergence rate assertions under source conditions and converse results are formulated. We note in passing that this paper was written by Mikhail Kokurin junior, who was supported by a scholarship of the President of the Russian Federation for young scientists and postgraduate students.
    0 references
    abstract Cauchy problem
    0 references
    ill-posed problem
    0 references
    Hilbert space
    0 references
    finite difference methods
    0 references
    quasi-reversibity method
    0 references
    convergence
    0 references
    converse theorems
    0 references
    0 references
    0 references

    Identifiers

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