Über das Divergenzverhalten des Intervall-Newton-Verfahrens. (On the divergence behaviour of the interval-Newton-method) (Q1179547)

Consider a function \(f: D\subseteq \mathbb{R}^ n\to\mathbb{R}^ n\). Let \(N\) denote the interval-Newton-operator for \(f\). Define an interval-iteration \([x]^{k+1}:=N[x]^ k\cap[x]^ k\). Under certain conditions it holds: If \(f\) has a zero \(x^*\) in the starting interval \([x]^ 0\) then the sequence \([x]^ 0,[x]^ 1,[x]^ 2,\dots\) converges quadratically towards \(x^*\). If on the other hand \(f\) has no zero in \([x]^ 0\) then after a finite number of steps \([x]^{k+1}=0\), i.e. the iteration breaks down, it ``diverges''. It is shown that in a certain sense this divergence behavior is quadratic. The latter property may be of relevance for instance for exclusion tests in the search of zeros of \(f\).