A unifying theorem for Newton's method on spaces with a convergence structure (Q617655)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A unifying theorem for Newton's method on spaces with a convergence structure
scientific article

    Statements

    A unifying theorem for Newton's method on spaces with a convergence structure (English)
    0 references
    0 references
    0 references
    21 January 2011
    0 references
    The article deals with Newton-like iteration for nonlinear equations \(F(x) = 0\) with operators \(F:\;{\mathfrak X} \to {\mathfrak X}\), where \({\mathfrak X}\) is a Banach space with a convergence structure \(({\mathfrak X},{\mathcal V},{\mathcal W})\). The authors mean the term ``a Banach space with a convergence structure'' as a triple \(({\mathfrak X},{\mathcal V},{\mathcal W})\) such that (1) \(({\mathfrak X},\|\cdot\|)\) is a real Banach space; (2) \(({\mathcal V},{\mathcal C},\|\cdot\|_{\mathcal V})\) is a real Banach space that is partially ordered by the closed convex cone \({\mathcal C}\) and the norm \(\|\cdot\|_{\mathcal V}\) is monotone; (3) \({\mathcal W}\) is a closed convex cone in \({\mathfrak X} \times {\mathcal V}\) such that \(\{0\} \times {\mathcal C} \subset {\mathcal W} \subset {\mathfrak X} \times {\mathcal C}\); (4) The following map \(/\cdot/: {\mathcal D} \to {\mathcal C}\), \(/x/ = \inf \;\{p \in {\mathcal C}:\;(x,p) \in {\mathcal W}\}\), where \({\mathcal D} = \{x \in {\mathfrak X}:\;\exists \, p \in {\mathcal C}, (x,p) \in {\mathcal W}\}\) is well defined; (5) For every \(x \in {\mathcal D}\) the inequality \(\|x\| \leq \|/x/\|_{\mathcal V}\) holds. The iteration under consideration is \[ x_0 = 0, \quad x_{n+1} = x_n + F'(x_n)^*(-F(x_n), \] where \(A^*\) (a right inverse) for \(A \in {\mathcal L}({\mathfrak X})\) is defined with the relation \[ A^*y = z \;\Leftrightarrow \;z \in T^\infty(0), \;\;T(x) = (I - A)x + y \;\Leftrightarrow \;z = \sum_{j=0}^\infty (I - A)^jy. \] The authors prove two analogues of the classical Kantorovich theorem about the semilocal convergence and illustrate these results with examples. It should be noted that in the article there are some vague places and it is written in a confusing manner.
    0 references
    0 references
    0 references
    0 references
    0 references
    Newton-Kantorovich method
    0 references
    convergence structure
    0 references
    Kantorovich hypothesis
    0 references
    nonlinear operator equations
    0 references
    numerical examples
    0 references
    Banach space
    0 references
    partially ordered
    0 references
    convex cone
    0 references
    0 references