Optimality of the Lax-Wendroff condition

From MaRDI portal





Let \(M_ n\) be the algebra of \(n\times n\) complex-valued matrices, F a subset of \(M_ n\), \(| \cdot |\) a vector norm on \(M_ n\). F is said stable if exists a constant \(K>0\) such that for every \(A\in F,| A^ k| \leq K\), \(k=1,2,...,\); the norm is stable if its unit ball, \(\{A| \quad | A| \leq 1\}\), is a stable set, and it is spectrally dominant if \(| A| \geq \rho(A)\), (spectral radius of A) for every \(A\in M_ n\). \textit{P. D. Lax} and \textit{B. Wendroff} [Comm. Pure Appl. Math. 17, 381-398 (1964; Zbl 0233.65050)] showed that the numerical radius, defined by \(r(A)=\max_{n}| \frac{(An,n)}{(n,n)}|\), is a stable norm, i.e., the set \(\{A| r(A)\leq 1\}\) is stable. In this paper the authors, imposing certain conditions on F, prove that any spectrally dominant vector norm which is invariant under similarity by isometries dominates the numerical radius, what means that the Lax-Wendroff stability condition \(r(\cdot)\leq 1\) defines a maximal isometrically invariant stable set.











This page was built for publication: Optimality of the Lax-Wendroff condition

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q788070)