\(H^{\infty}\) functional calculus and square function estimates for Ritt operators (Q2256069): Difference between revisions

A linear operator on a Banach space is said to be a \textit{Ritt operator} if the sequences \(\{T^n\}\) and \(\{n(T^n - T^{n-1})\}\) are bounded in \(\mathcal B(X)\). Note that any strict contraction is a rather special example of Ritt operator. The following characterization can be used to generate examples of such operators (Lemma 2.1): If \(B_{\gamma}\) denotes the interior of the convex hull of \(1\) and the disc \(\mathbb D_{\sin \gamma}\) (Stoltz domain) for \(\gamma \in (0, \pi/2)\), then \(T\) is a Ritt operator if and only if there exists \(\alpha \in (0, \pi/2)\) such that \(\sigma(T) \subseteq \overline{B}_{\alpha}\), and, for any \(\beta \in (\alpha, \pi/2)\), the set \(\{(\lambda -1)(T-\lambda)^{-1} : \lambda \notin \overline{B}_{\beta}\}\) is bounded. The last characterization also connects the theory of Ritt operators with that of sectorial operators. Indeed, if \(T\) is a Ritt operator (of type \(\alpha\)), then \(I-T\) is a sectorial operator of type \(\alpha\). This characterization is, then used to construct an \(H^{\infty}(B_{\gamma})\) (for short, \(H^{\infty}\)) functional calculus for Ritt operators of type \(\alpha\): For any \(\gamma \in (0, \pi/2)\), let \(H^{\infty}_0\) be the space of bounded holomorphic functions \(\phi : B_{\gamma} \rightarrow \mathbb C\) for which there exist real numbers \(s, c > 0\) such that \[ |\phi(\lambda)| \leq c|1-\lambda|^s,\;\lambda \in B_{\gamma}. \] The functional calculus is given by \[ \phi(T) = \frac{1}{2 \pi i}\int_{\partial B_{\beta}}\phi(\lambda)(T-\lambda)^{-1}d\lambda, \] where \(\beta \in (\alpha, \gamma)\) and the boundary \(\partial B_{\beta}\) carries counter-clockwise orientation. Note that \(\phi \mapsto \phi(T)\) defines an algebra homomorphism from \(H^{\infty}_0\) into \(\mathcal B(X)\), which can be made unital by setting \(\psi(T) = c I + \phi(T)\) for \(\psi = c + \phi\) and scalar \(c\). In particular, this functional calculus is applicable to polynomials. It is shown in Proposition 2.5 that polynomial functional calculus determines bounded \(H^{\infty}\) functional calculus in the following sense: A Ritt operator \(T\) has a bounded \(H^{\infty}\) functional calculus if and only if there exists a constant \(K \geq 1\) such that, for any polynomial \(\phi\), \(\|\phi(T)\| \leq K \|\phi\|_{\infty, B_{\gamma}}\). In particular, any Ritt operator with bounded \(H^{\infty}\) functional calculus is polynomially bounded. A partial converse of this for some special Ritt operators is presented in Proposition 7.7. Among many results, we discuss below two results of the paper under review. The first one is a transfer principle from sectorial operators to Ritt operators: \(T\) admits a bounded \(H^{\infty}(B_{\gamma})\) functional calculus for some \(\gamma \in (0, \pi/2)\) if and only \(I-T\) admits a bounded \(H^{\infty}(\Sigma_{\theta})\) functional calculus for some \(\theta \in (0, \pi/2)\), where \(\Sigma_{\theta}\) is the open sector \(\{z \in \mathbb C : z \neq 0, |\text{Arg}(z)| < \theta\}\). The second result is a single operator analogue of the Cowling-Doust-McIntosh-Yagi characterization of bounded \(H^{\infty}\) functional calculus on \(L^p\)-spaces by the boundedness of certain Littlewood-Paley-Stein square functions [\textit{M. Cowling} et al., J. Aust. Math. Soc., Ser. A 60, No. 1, 51--89 (1996; Zbl 0853.47010)]. For the precise statement, see Theorem 7.3 and Corollary 7.5 of Section 7.