Optimal Enclosure of a Generalized Zero Set of a Function Strip

From MaRDI portal





This paper is concerned with a class of iterative methods for solving \(g(x)=0\), where \(g: D\subseteq {\mathbb{R}}\) \(n\to {\mathbb{R}}^ n.\) The author investigates the case where the function g is not exactly known. The only information about g is that g can be bounded by known functions \b{g}: \(D\to {\mathbb{R}}^ n \)and \(\bar g:\) \(D\to {\mathbb{R}}^ n \)with \(\underline g(x)\leq g(x)\leq \bar g(x)\) for all \(x\in D\). Let I(\({\mathbb{R}}^ n\)) denote the set of compact intervals in \({\mathbb{R}}^ n.\) Define the function \(G: D\to I({\mathbb{R}}^ n\)) with \(G(x):=[\underline g(x),\bar g(x)]\) and the zero set X \(*:=\{x| x\in D\), \(0\in G(x)\}.\) The zeros of g are certainly contained in the zero set X *. For some fixed matrix \(b\in {\mathbb{R}}^{n\times n}\) define the function \(F: D\to I({\mathbb{R}}^ n\)) with \(F(x):=x-b\cdot G(x)\) and the fixed point set \(X^{**}:=\{x| x\in D,\quad x\in F(x)\}.\) Obviously X \(*\subseteq X^{**}\). Let \(I(D):=\{A| \quad A\in I({\mathbb{R}}\quad n),\quad A\subseteq D\}\) and \(F(x)=[\underline f(x),\bar f(x)\}\) for \(x\in D\). The author defines and discusses the operator Q: I(D)\(\to I({\mathbb{R}}^ n\)) with \(Q([\underline x,\bar x]):=[\underline f(\underline x),\bar f(\bar x)].\) Especially \(X\in I(D)\) with \(Q(X)=X\) is of interest. The iteration \(X_ 0\in I(D)\), \(X_{k+1}:=Q(X_ k)\cap X_ k\) for \(k=0,1,2,..\). defines a sequence of intervals in \({\mathbb{R}}^ n.\) Under certain conditions this sequence converges to a limit interval which optimally encloses the fixed point set \(X^{**}\).











This page was built for publication: Optimal Enclosure of a Generalized Zero Set of a Function Strip

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3776509)