Boundedness properties of the maximal operator in a nonsymmetric inverse Gaussian setting

Tommaso Bruno, Valentina Casarino, Paolo Ciatti, Peter Sjögren · 2025

Read Boundedness properties of the maximal operator in a nonsymmetric inverse Gaussian setting by Tommaso Bruno, Valentina Casarino, Paolo Ciatti, Peter Sjögren (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-2501-17517.txt.

Introduction

Tommaso Bruno, Valentina Casarino, Paolo Ciatti, and Peter Sjögren proved bounds for a maximal operator in a broadened inverse Gaussian setting in 2025. The operator is tied to a nonsymmetric heat-like flow. Such bounds say how strongly the operator can stretch functions in a given norm.

They show the operator is bounded for a wide range of exponents and is of weak type at the low end. For short times, the proof uses a forbidden zones method first made in the Gaussian case. For large times, that method is not enough, so the authors add new tools built for this wider setting.

Catalog record

CREATORTommaso Bruno, Valentina Casarino, Paolo Ciatti, Peter Sjögren
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-17517

Browse the full library · About the Tower of Babel

Full text

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

Boundedness properties of the maximal operator in a nonsymmetric inverse Gaussian setting Thanks: The first three authors are members of the Gruppo Nazionale per l’Analisi Matematica, la Probabilità e le loro Applicazioni (GNAMPA) of the Istituto Nazionale di Alta Matematica (INdAM) and were partially supported by the INdAM–GNAMPA Project “ L p L^{p} estimates for singular integrals in nondoubling settings” (CUP E53C23001670001). This research was carried out while the fourth author was a visiting scientist at the University of Padova, Italy, and he is grateful for its hospitality. Tommaso Bruno, Valentina Casarino, Paolo Ciatti, Peter Sjögren Abstract . We introduce a generalized inverse Gaussian setting and consider the maximal operator associated with the natural analogue of a nonsymmetric Ornstein–Uhlenbeck semigroup. We prove that it is bounded on L p L^{p} when p ∈ ( 1 , ∞ ] p\in(1,\infty] and that it is of weak type ( 1 , 1 ) (1,1) , with respect to the relevant measure. For small values of the time parameter t t , the proof hinges on the “forbidden zones” method previously introduced in the Gaussian context. But for large times the proof requires new tools. 2000 Mathematics Subject Classification 42B25, 47D03. 1. Introduction Alongside with the Gaussian framework, the Euclidean setting endowed with the absolutely continuous measure whose density is the reciprocal of a Gaussian has acquired independent interest in the last decade. It is nowadays known as the inverse Gaussian setting. Its introduction in the realm of harmonic analysis dates back to F. Salogni [ 24 ] , who introduced the operator \mathcal{A}=\tfrac{1}{2}\,\Delta+\langle x,\nabla\rangle as an (essentially) self-adjoint operator in L^{2}(\mathbb{R}^{n},d\gamma_{-1}) . Here d\gamma_{-1} stands for the inverse Gaussian measure d\gamma_{-1}(x)=\pi^{n/2}\,e^{|x|^{2}}\,dx. In her PhD thesis [ 24 ] , Salogni proved the weak type (1,1) of the maximal operator associated to the semigroup generated by \mathcal{A} . A few years later, T. Bruno and P. Sjögren [ 6 , 7 ] (see also [ 5 ] ) studied Riesz transforms and Hardy spaces for \mathcal{A} . Several other contributions to harmonic analysis in this context appeared more recently, see [ 1 , 2 , 3 , 4 , 20 ] . While this development of the inverse Gaussian setting was going on, V. Casarino, P. Ciatti and P. Sjögren started investigating several classical problems related to the semigroup generated by the generalized Ornstein–Uhlenbeck operator (1.1) \mathcal{L}^{Q,B}=\tfrac{1}{2}\,\mathrm{tr}\big(Q\nabla^{2}\big)+\langle Bx,\nabla\rangle in the Gaussian setting. Here \nabla indicates the gradient, \nabla^{2} the Hessian, and Q and B are two real n\times n matrices called covariance and drift, respectively, satisfying (H1) Q is symmetric and positive definite; (H2) all the eigenvalues of B have negative real parts. We refer the reader to [ 8 , 9 , 10 , 11 , 12 , 13 , 15 ] and the brief overview [ 14 ] . The semigroup generated by \mathcal{L}^{Q,B} has the Gaussian probability measure (1.2) d\gamma_{\infty}(x)=(2\pi)^{-\frac{n}{2}}(\text{det}\,Q_{\infty})^{-\frac{1}{2}}\exp\big({-\tfrac{1}{2}\langle Q_{\infty}^{-1}x,x\rangle}\big)\,dx as invariant measure. Here Q_{\infty} is a certain symmetric and positive definite matrix whose precise definition in terms of Q and B will be given in Section 2 . In some particular cases, for instance when Q and -B coincide with the identity matrix I_{n} , the operator \mathcal{L}^{Q,B} is essentially self-adjoint in L^{2}(d\gamma_{\infty}) , but in general it is not even symmetric. Thus many classical problems, like the boundedness of singular integral operators associated with \mathcal{L}^{Q,B} , give rise to interesting and nontrivial questions. New techniques and new ideas are required, strong enough to overcome the lack of symmetry. In this article, we combine the two approaches by considering a generalized version of the inverse Gaussian setting studied in [ 24 , 6 , 7 ] via some of the techniques developed in [ 8 , 9 , 10 , 11 , 12 , 13 , 15 ] and others which are new. This may be seen as the starting-point of a program whose goal is to develop an analogous theory in a nonsymmetric inverse Gaussian setting. We shall replace the density of d\gamma_{\infty} by its reciprocal, i.e., \mathbb{R}^{n} will be equipped with the inverse Gaussian measure d{\gamma_{-\infty}}(x)=(2\pi)^{\frac{n}{2}}(\text{det}\,Q_{\infty})^{\frac{1}{2}}\exp\big({\tfrac{1}{2}\langle Q_{\infty}^{-1}x,x\rangle}\big)\,dx. Like d\gamma_{\infty} , this measure is obviously locally doubling but not globally doubling, but in contrast to the probability measure d\gamma_{\infty} , it has superexponential growth at infinity. In this setting, the role which was played by the Ornstein–Uhlenbeck operator in (\mathbb{R}^{n},d\gamma_{\infty}) is now played by the so-called inverse Ornstein–Uhlenbeck operator (1.3) \mathcal{A}^{Q,B}=\tfrac{1}{2}\,\mathrm{tr}\big(Q\nabla^{2}\big)-\langle Bx,\nabla\rangle. Here B and Q are exactly the covariance and drift matrices inherited from the Ornstein–Uhlenbeck setting and satisfying (H1) and (H2). Notice that d\gamma_{-1} and \mathcal{A} , considered in [ 1 , 2 , 3 , 4 , 6 , 7 , 20 , 24 ] , are a special case of d\gamma_{-\infty} and \mathcal{A}^{Q,B} , corresponding to the choice Q=-B=I_{n} . The semigroup generated by \mathcal{A}^{Q,B} is \mathcal{H}_{t}^{UO}:=e^{t\mathcal{A}^{Q,B}}\,,\qquad t>0, (here “ UO ” has the scope of emphasizing the contrast with respect to “ OU ” which will be used for Ornstein–Uhlenbeck). While in the Gaussian context d\gamma_{\infty} is invariant under the action of the Ornstein–Uhlenbeck semigroup, and therefore chosen as the substitute for Lebesgue measure, d\gamma_{-\infty} is not invariant under the action of (\mathcal{H}_{t}^{UO}) . In fact, there are no invariant measures for (\mathcal{H}_{t}^{UO}) ; see, e.g., [ 21 ] or [ 19 ] . Nonetheless, d\gamma_{-\infty} appears to be the natural measure for (\mathcal{H}_{t}^{UO}) , because whenever QB^{*}=BQ the operators \mathcal{H}_{t}^{UO} are symmetric in L^{2}(\mathbb{R}^{n},d\gamma_{-\infty}) (see Remark 2.3 ). As we shall see in Proposition 2.2 , each operator of this semigroup is an integral operator, with a kernel K_{t}^{UO} with respect to d\gamma_{-\infty} . In this paper we study the boundedness of the maximal operator associated to the semigroup (\mathcal{H}_{t}^{UO})_{t>0} , proving that it is of weak type (1,1) and of strong type (p,p) for all 1<p\leq\infty , with respect to d\gamma_{-\infty} . This extends a similar result in [ 24 ] , proved under the assumption Q=-B=I_{n} . Let us note that our results appear to be the first of their kind for nonsymmetric operators on manifolds with superexponential volume growth. As is standard by now, the proof distinguishes between the local and global parts of the kernel K_{t}^{UO} of \mathcal{H}_{t}^{UO} . Here local and global mean that |x-u|\leq 1/(1+|x|) and |x-u|>1/(1+|x|) , respectively, x,u being the two arguments of the kernel. Beyond this distinction that dates back to [ 23 , 25 , 17 ] , the techniques in [ 24 ] rely on the spectral resolution of the self-adjoint operator \mathcal{A}=\mathcal{A}^{I_{n},-I_{n}} and seem no longer applicable in a nonsymmetric context. Fortunately, large parts of the machinery developed in [ 8 , 9 , 10 , 11 , 12 , 13 , 15 ] to study Gaussian harmonic analysis in a nonsymmetric setting can be transferred to the inverse setting, and are useful to treat the local part of the maximal function and its global part for t\in(0,1] . The global part for t\geq 1 is more delicate and requires new tools. Structure of the paper In Section 2 we recall some basic facts concerning the Ornstein–Uhlenbeck setting. We also compute an explicit expression for the inverse Mehler kernel K_{t}^{UO} and discuss its relationship with the Gaussian Mehler kernel.

Read the complete text