Hörmander-Mikhlin type theorem on non-commutative spaces

Rauan Akylzhanov, Michael Ruzhansky, Kanat Tulenov · 2025

Read Hörmander-Mikhlin type theorem on non-commutative spaces by Rauan Akylzhanov, Michael Ruzhansky, Kanat Tulenov (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-01240.txt.

Introduction

Rauan Akylzhanov, Michael Ruzhansky, and Kanat Tulenov carried a old-style Fourier multiplier theorem into non-commutative settings in a 2025 paper. The Hörmander–Mikhlin theorem tells analysts which rate multipliers act boundedly on function spaces, a workhorse of harmonic analysis. Their version develops a Fourier formalism on non-commutative spaces, where functions are replaced by functions and multiplication no longer commutes. Then proves the multiplier theorem twice over: for locally compact Kac groups and for semi-finite von Neumann algebras.

In the simplest case the result reproduces a sharp form of the old-style theorem. Applications to evolution math rules in the non-commutative setting follow. Such theorems underpin quantum probability and non-commutative shapes, where old-style Fourier tools must be rebuilt from scratch.

Catalog record

CREATORRauan Akylzhanov, Michael Ruzhansky, Kanat Tulenov
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-01240

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Hörmander-Mikhlin type theorem on non-commutative spaces Rauan Akylzhanov, Michael Ruzhansky, Kanat Tulenov Abstract . In this paper, we introduce a Fourier-type formalism on non-commutative spaces. As a result, we obtain two versions of Hörmander-Mikhlin L p L^{p} -multiplier theorem: on general von Neumann algebras and semi-finite von Neumann algebras respectively. We develop elements of noncommutative Littlewood-Paley theory to connect with classical symbolic estimates. In the particular case of ℝ \mathbb{R} , we explicitly show how to recover a sharp version of the classical Hörmander L p L^{p} -multiplier theorem, which was obtained by Grafakos and Slavíková in [ 14 ] . Finally, we apply the main results to obtain time-asymptotics of abstract Klein-Gordon equation on semi-finite von Neumann algebras. 2020 Mathematics Subject Classification 46L51, 47L25, 42B15, 43A85, 43A15, 42A38, 43A32, 35S05. 1. Introduction Let \sigma be a bounded function on \mathbb{R}^{n}. Then for any function f in the Schwartz space \mathcal{S}(\mathbb{R}^{n}) define a linear operator A_{\sigma} on \mathcal{S}(\mathbb{R}^{n}) as follows (1.1) A_{\sigma}f(\xi)=\int_{\mathbb{R}^{n}}\widehat{f}(\eta)\sigma(\eta)e^{2\pi i(\xi,\eta)}d\eta, where \widehat{f} is the Fourier transform \widehat{f}(\xi)=\int_{\mathbb{R}^{n}}f(\eta)e^{-2\pi i(\xi,\eta)}d\eta,\quad\xi\in\mathbb{R}^{n}, of the function f, and (\xi,\eta) denotes the standard inner product of vectors \xi and \eta in \mathbb{R}^{n}. One of the main problems in harmonic analysis is to find a sufficient condition on \sigma that guarantees that the operator A_{\sigma} is bounded on L^{p}(\mathbb{R}^{n}) for a given 1<p<\infty. In this case, the operator A_{\sigma} will be called Fourier multiplier with the symbol \sigma. In [ 30 ] , Mikhlin showed that if the condition |\partial^{\alpha}\sigma(\xi)|\leq C_{\alpha}|\xi|^{-|\alpha|},\quad 0\neq\xi\in\mathbb{R}^{n}, holds for any multi-indices \alpha such that |\alpha|\leq[\frac{n}{2}]+1, then the operator A_{\sigma} has a bounded extension from L^{p}(\mathbb{R}^{n}) into itself for any 1<p<\infty. Later, Hörmander extended in [ 17 ] Mikhlin’s result. Indeed, Hörmander’s theorem says that if (I-\Delta_{\mathbb{R}^{n}})^{s/2} is the operator given on the Fourier transform by multiplication by (1+4\pi^{2}|\xi|^{2})^{s/2} for some s>0 and let \Psi be a Schwartz function on \mathbb{R}^{n} whose Fourier transform is supported in the annulus \frac{1}{2}<|\xi|<2 and satisfies \sum\limits_{j\in\mathbb{Z}}\widehat{\Psi}\left(2^{-j}\xi\right)=1,\xi\neq 0 . Moreover, if for some 1\leq r\leq 2 and s>n/r, the function \sigma satisfies the condition (1.2) \sup_{j\in\mathbb{Z}}\left\|(I-\Delta_{\mathbb{R}^{n}})^{\frac{s}{2}}\left[\widehat{\Psi}\sigma(2^{j}\cdot)\right]\right\|_{L^{\frac{n}{s},\,1}\left(\mathbb{R}^{n}\right)}<\infty. This is a Littlewood–Paley formulation: the symbol is localised to dyadic annuli via \sigma(2^{j}\cdot) and a smooth \Psi supported in \tfrac{1}{2}<|\xi|<2 with \sum_{j}\widehat{\Psi}(2^{-j}\xi)=1 ; it is the standard way to verify multiplier bounds in the Euclidean setting. Then A_{\sigma} has a bounded extension from L^{p}(\mathbb{R}^{n}) into itself for any 1<p<\infty. In [ 4 ] , Caldéron and Torchinsky obtained that if condition ( 1.2 ) holds with p satisfying \left|\frac{1}{p}-\frac{1}{2}\right|<\frac{s}{n} and \left|\frac{1}{p}-\frac{1}{2}\right|=\frac{1}{r}, then A_{\sigma} is bounded from L^{p}(\mathbb{R}^{n}) into itself. Very recently, Grafakos and Slavíková in [ 14 ] improved Hörmander’s theorem and the result of Caldéron and Torchinsky by replacing the norm in ( 1.2 ) with the Lorentz space L^{\frac{n}{s},1}\left(\mathbb{R}^{n}\right) -quasi-norm. Their main result is given as follows: Theorem 1.1 . [ 14 , Theorem 1.1] . Let \Psi be a Schwartz function on \mathbb{R}^{n} whose Fourier transform is supported in the annulus \frac{1}{2}<|\xi|<2 and satisfies \sum\limits_{j\in\mathbb{Z}}\widehat{\Psi}\left(2^{-j}\xi\right)=1,\quad\xi\neq 0 . Let 1<p<\infty and let s\in(0,n),n\in\mathbb{N} , satisfy \left|\frac{1}{p}-\frac{1}{2}\right|<\frac{s}{n}. Then for all functions f\in\mathcal{S}\left(\mathbb{R}^{n}\right) we have the following a priori estimate (1.3) \|A_{\sigma}f\|_{L^{p}(\mathbb{R}^{n})}\leq C\sup_{j\in\mathbb{Z}}\left\|(I-\Delta_{\mathbb{R}^{n}})^{\frac{s}{2}}\left[\widehat{\Psi}\sigma(2^{j}\cdot)\right]\right\|_{L^{\frac{n}{s},1}\left(\mathbb{R}^{n}\right)}\|f\|_{L^{p}(\mathbb{R}^{n})} with some constant C>0. The main aim of this paper is to obtain variants of the Hörmander L^{p} -multiplier theorem on general von Neumann algebras. Let \M be a von Neumann algebra and let a linear operator \D\colon L^{2}(\M)\rightarrow L^{2}(\M) be affiliated with the dual von Neumann algebra \Mh . We prove the L^{p}(\M)\to L^{p}(\M) boundedness of Fourier multipliers in two complementary forms: a global estimate (operator norm on \widehat{\M} ) and a local estimate (Littlewood–Paley decomposition, fibre-wise norms), the latter recovering the classical annulus condition of Grafakos–Slavíková [ 14 ] on \mathbb{R}^{n} (Example 6.1.1 ). The global form (Section 5 , Theorem 5.2 ) takes the form: (1.4) \|A\|_{L^{p}(\M)\rightarrow L^{p}(\M)}\leq\left\|\Dh^{-\beta}\right\|_{L^{\infty}(\Mh)}\left\|\Dh^{\beta}A\right\|_{L^{r}(\Mh)},\quad\frac{1}{r}=2\left\lvert\frac{1}{p}-\frac{1}{2}\right\rvert,\quad 1<p<\infty. The local (Littlewood–Paley) form (Section 6 , Theorem 6.1 ) takes the following form under a direct-integral decomposition \widehat{\M}=\int_{\widehat{Z}}^{\oplus}\Mh_{\lambda}\,d\mu(\lambda) and a Littlewood–Paley decomposition over \widehat{Z} ; here the T_{j}\colon\widehat{Z}\to\widehat{Z} are the measurable maps of that decomposition (so T_{j}^{-1}(E_{j})=E_{0} for all j and they have uniformly bounded Radon–Nikodým derivatives; see Section 6 ). (1.5) \|A\|_{L^{p}(\M)\rightarrow L^{p}(\M)}\lesssim\sup_{j\in\mathbb{Z}}\Bigl\|\bigl\|\LL(\lambda)A(T_{j}(\lambda))\bigr\|_{L^{r,\infty}(\Mh_{\lambda})}\Bigr\|_{L^{r,1}(\widehat{Z})},\quad\frac{1}{r}=2\left\lvert\frac{1}{p}-\frac{1}{2}\right\rvert,\quad 1<p<\infty. Here, we choose \beta>0 such that \left\|\Dh^{-\beta}\right\|_{L^{\infty}(\Mh)} and the relevant L^{r}(\Mh) -norms are finite; the local form requires the LP decomposition setup of Section 6 . As a particular example, we recover Theorem 1.1 of [ 14 ] in Example 6.1.1 . For an \LL -wave equation \frac{\partial^{2}{u}}{{\partial t^{2}}}+\LL u=0,\partial u=u_{1} , we prove a time-decay estimate for the propagator u(t)=\frac{\sin(t\cdot\sqrt{\LL})}{\sqrt{\LL}}u_{1}+\cos(t\sqrt{\LL})u_{0} . \LL^{-\beta}\frac{\sin t\sqrt{\LL}}{\sqrt{\LL}} (Theorem 7.2 ): under a Weyl law \tau(E_{(\kappa,s)}(\LL))\cong s^{\alpha} and \frac{1}{r}\geq\frac{\beta}{\alpha} , 0<\gamma\leq 2 , (1.6) \|u(t)\|_{L^{q}(\M)}\lesssim(C_{\gamma,p,q,\beta,\kappa,\LL}\cdot t^{1-\gamma\beta-\frac{\alpha}{\beta}\gamma\frac{1}{r}}+1)(\|\LL^{\beta}u_{1}\|_{L^{p}(\M)}+\|u_{0}\|_{L^{p}(\M)}). The structure of the paper is as follows. Section 2 introduces preliminaries on noncommutative L^{p} and Lorentz spaces and the notion of Fourier structure on von Neumann algebras; it is used in all subsequent sections. Section 3 develops Hausdorff–Young and Paley-type inequalities under the Fourier structure; these feed into Section 4 (Hardy–Littlewood) and into Section 5 (multiplier theorems). Section 4 establishes Hardy–Littlewood-type inequalities involving \D^{-\beta} and the dual theorem (Theorem 4.3 ); the latter is the key step in the Hörmander–Mihlin type proofs in Section 5 . Section 5 proves the main Hörmander–Mihlin multiplier theorems (off-diagonal L^{p} – L^{q} and diagonal semifinite Lorentz version); it relies on Sections 2 – 4 and gives operator-norm bounds. Section 6 develops a Littlewood–Paley framework (reduction theory, direct integrals, decompositions over the spectrum of an abelian subalgebra) so that multiplier boundedness can be stated in symbolic/fibre form (Theorem 6.1 ); it uses the same Fourier structure and multiplier theorems, and Example 6.1.1 shows that this formulation recovers the classical Grafakos–Slavíková condition [ 14 ] on \mathbb{R}^{n} . Section 7 presents applications to evolution equations (heat and wave propagators), using the multiplier theorems and Lorentz norms; it recovers results from [ 14 , 39 , 26 , 41 ] . We also note that the L^{p} -boundedness of Fourier multipliers can be extended to the range 0<p<\infty in the context of Hardy spaces; see the results on general graded groups in [ 19 ] and references therein. 2. Preliminaries 2.1. Classical L^{p} and L^{p,q} spaces For 0<p\leq\infty and \mathbb{R}_{+}:=(0,\infty), the space L^{p}(\mathbb{R}_{+}) refers to the L^{p} -space of equivalence classes of p -integrable functions, defined almost everywhere, and L^{\infty}(\mathbb{R}_{+}) denotes the space of essentially bounded functions on the positive half-line \mathbb{R}_{+}. For 0<p,q\leq\infty, the Lorentz space L^{p,q}(\mathbb{R}_{+}) consists of all complex-valued measurable functions f on \mathbb{R}_{+} for which the following quasi-norm is finite: \|f\|_{L^{p,q}(\mathbb{R}_{+})}=\left\{\begin{array}[]{rcl}\left(\int\limits_{\mathbb{R}_{+}}\left(t^{\frac{1}{p}}f^{*}(t)\right)^{q}\frac{dt}{t}\right)^{\frac{1}{q}},&\mbox{for}&q<\infty;\\ \sup\limits_{t>0}t^{\frac{1}{p}}f^{*}(t),&\mbox{for}&q=\infty.\end{array}\right. Here, f^{*} denotes the decreasing rearrangement of the function f. The quasi-norm in L^{p,\infty}(\mathbb{R}_{+}) is also given as follows (2.1) \|f\|^{p}_{L^{p,\infty}(\mathbb{R}_{+})}=\sup\limits_{t>0}t^{p}m\{s>0:|f(s)|>t\}, where m is the Lebesgue measure on \mathbb{R}_{+}. For more information on these spaces and their properties, we refer the reader to [ 13 ] . 2.2. Von Neumann algebras and \tau -measurable operators Let \HH be a complex Hilbert space, and let B(\HH) denote the algebra of all bounded linear operators on \HH. Consider a semi-finite von Neumann algebra \M\subset B(\HH) equipped with a faithful, normal, semi-finite trace \tau. The pair (\M,\tau) is referred to as a non-commutative measure space . A closed, densely defined operator x on \HH with domain \mathrm{dom}(x), is called affiliated with \M if u^{\ast}xu=x for every unitary

Read the complete text