Trace type Orlicz spaces and analysis of Orlicz spaces by Lebesgue exponents
Albin Petersson · 2025
Read Trace type Orlicz spaces and analysis of Orlicz spaces by Lebesgue exponents by Albin Petersson (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-2505-21764.txt.
Introduction
Albin Petersson analyzes Orlicz spaces, a flexible family of function spaces, in a 2025 math paper. Each such space is built from a Young function that sets how large values are punished. Two Lebesgue exponents summarize where the space sits among classic power spaces.
Petersson shows that between those bounds one can always swap in an equivalent Young function with exponents squeezed to a chosen range. That swap sharpens known inclusions between the Orlicz space and its Lebesgue neighbors. For trace type spaces, he ties a key inclusion to a single inequality, with the reverse inclusion matching the reversed inequality. The work tightens the map of these spaces.
Catalog record
| CREATOR | Albin Petersson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 76 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-21764
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Trace type Orlicz spaces and analysis of Orlicz spaces by Lebesgue exponents Albin Petersson Abstract . In the paper, we analyze the Lebesgue exponents p Φ p_{\Phi} and q Φ q_{\Phi} , and show that for any p Φ < p < ∞ p_{\Phi}<p<\infty and 1 < q < q Φ 1<q<q_{\Phi} , there exists an equivalent Young function Ψ \Psi with p < p Ψ < ∞ p<p_{\Psi}<\infty and 1 < q Ψ < q 1<q_{\Psi}<q . This type of construction is used to improve upon the inclusions L p Φ ∩ L q Φ ⊆ L Φ ⊆ L p Φ + L q Φ L^{p_{\Phi}}\cap L^{q_{\Phi}}\subseteq L^{\Phi}\subseteq L^{p_{\Phi}}+L^{q_{\Phi}} . For trace type Orlicz spaces L Φ , Φ L^{\Phi,\Phi} , we find that when Φ ∈ Δ 2 \Phi\in\Delta_{2} , we have L Φ , Φ ⊆ L Φ L^{\Phi,\Phi}\subseteq L^{\Phi} if and only if Φ ( ‖ f ‖ L Φ ) ≤ C ρ Φ ( f ) \Phi(\|f\|_{L^{\Phi}})\leq C\rho_{\Phi}(f) for all f ∈ L Φ f\in L^{\Phi} , and the reverse inclusion is equivalent to the reversed inequality. 0. Introduction A fundamental setting for mathematics in numerous fields is the Lebesgue space L^{p} , 1\leq p\leq\infty (with important special cases p=1,2 , or \infty ). However, some problems require the examination of non-homogeneous functionals which do not neatly slot into any Lebesgue space. One example from statistical physics is the entropy E of a probability density function f given by E(f)=-\int f(\xi)\log f(\xi)\,d\xi, which is not well-defined for all f\in L^{1} . To solve this problem, one turns to the spaces introduced by W. Orlicz in [ 10 ] , which we now call Orlicz spaces. These are denoted by L^{\Phi} , where \Phi is a Young function, a certain type of convex function. The special case \Phi(t)=t^{p} , 1\leq p<\infty , gives L^{\Phi}=L^{p} . By [ 6 ] , one can avoid the continuity problem for E above by replacing the classical statistical framework \langle L^{\infty},L^{1}\rangle with \langle L^{\Psi},L^{\Phi}\rangle , where \Phi(t)=t\log(1+t) and \Psi(t)=\cosh t-1 . Since the Orlicz space L^{\Phi} is completely determined by the Young function \Phi , it is clear that the analysis of Young functions and their properties is essential. This analysis is mainly done via certain indices, which provide an estimate of the growth rate of a Young function \Phi . We consider quantities such as the limits at the origin and infinity of \frac{t\Phi^{\prime}(t)}{\Phi(t)} and \frac{\ln\Phi(t)}{\ln t} . Such indices, and many more, appear in [ 7 , 8 ] . Much of the analysis of this paper revolves around the so-called Lebesgue exponents p_{\Phi} and q_{\Phi} , which appeared in [ 12 ] , and have subsequently been analyzed in, e.g., [ 3 ] . We will mainly focus on the case that p_{\Phi}<\infty , which is equivalent to \Phi fulfilling the (global) \Delta_{2} -condition (see Remark 2.4 ). In this paper, we examine the Lebesgue exponents closely and determine their most glaring limitations. Namely, we show how equivalent Young functions can be constructed with widely differing Lebesgue exponents. We also present some alternative methods of obtaining growth estimates of Young functions, and by extension, obtain inclusion properties for Orlicz spaces. Furthermore, we investigate mixed Orlicz spaces, which are useful in interpolation theory. Particularly, we investigate the trace type Orlicz spaces L^{\Phi,\Phi}({\mathbb{R}}^{2d}) , where the same Young function \Phi is used for the two d -dimensional variables, but separately, as opposed to L^{\Phi}({\mathbb{R}}^{2d}). In the case of Orlicz modulation spaces, as was shown in [ 13 , Corollary 2.5] , we have M^{\Phi}({\mathbb{R}}^{d})=M^{\Phi,\Phi}({\mathbb{R}}^{d}) , meaning that \|V_{\varphi}\cdot\|_{L^{\Phi}} is equivalent to \|V_{\varphi}\cdot\|_{L^{\Phi,\Phi}} , where V_{\varphi} denotes the short-time Fourier transform with window function \varphi . This equivalence does not hold in general. Inclusions between the spaces L^{\Phi,\Phi} and L^{\Phi} are shown to be linked to inequalities between the quantities \rho_{\Phi}(f)=\int\Phi(|f(x)|)\,dx and \Phi(\|f\|_{L^{\Phi}}) . We find that for \Phi\in\Delta_{2} , L^{\Phi,\Phi}\subseteq L^{\Phi} if and only if \rho_{\Phi}(f)\leq C\Phi(\|f\|_{L^{\Phi}}) for some constant C>0 , which in turn is equivalent to \Phi being submultiplicative. The reversed inclusion is equivalent to the reverse inequality holding, which in turn is equivalent to \Phi being supermultiplicative. Evidently, L^{\Phi}=L^{\Phi,\Phi} therefore holds if and only if C^{-1}\Phi(\|f\|_{L^{\Phi}})\leq\rho_{\Phi}(f)\leq C\Phi(\|f\|_{L^{\Phi}}) for some constant C>0 , which by [ 4 ] is also equivalent to L^{\Phi}=L^{p} for some p\geq 1 . The paper is organized as follows. In Section 1 , we introduce necessary notations and definitions. The aforementioned Lebesgue exponents are examined in Section 2 . Their limitations are described in Section 2.1 , where it is found that if p_{\Phi}<p<\infty and 1<q<q_{\Phi} , then there exists an equivalent Young function \Psi with p<p_{\Psi}<\infty and 1<q_{\Psi}<q ( Proposition 2.12 ). In Section 2.2 and Section 2.3 , we rectify these problems in different ways by focusing on shared properties for the equivalence class and ultimately obtain better growth estimates for \Phi ( Proposition 2.20 and Theorem 2.24 ). The relation between spaces L^{\Phi} and L^{\Phi,\Phi} is explored in Section 3 . Here, Section 3.1 is focused on the Young function modular \rho_{\Phi}(f)=\int\Phi(|f(x)|)\,dx , and Section 3.2 contains sufficient and necessary conditions on \Phi , \Phi_{1} , and \Phi_{2} for inclusions between L^{\Phi} and L^{\Phi_{1},\Phi_{2}} to hold. In particular, such results are stated in the case that \Phi=\Phi_{1}=\Phi_{2} ( Corollary 3.13 and Corollary 3.15 ). 1. Preliminaries By A\subseteq B , we mean that A is a (not necessarily proper) subset of B , and we will write A\subsetneq B to specify that A is a proper subset of B . We will use d exclusively to denote dimension, and we will deal only with functions whose domains are \mathbb{R}_{+}=\{\,t\in\mathbb{R}\,:\,t\geq 0\,\}, {\mathbb{R}}^{d} , or {\mathbb{R}}^{2d} . For p\in[1,\infty] , let L^{p}({\mathbb{R}}^{d})=L^{p} denote the usual Lebesgue space with norm \|f\|_{L^{p}({\mathbb{R}}^{d})}\equiv\|f\|_{L^{p}}\equiv\|f\|_{p}\equiv\begin{cases}\left(\int_{{\mathbb{R}}^{d}}|f(x)|^{p},dx\right)^{\frac{1}{p}},&p<\infty,\\[4.30554pt] \displaystyle{\esssup_{x\in{\mathbb{R}}^{d}}}|f(x)|,&p=\infty,\end{cases} where {f}:{{\mathbb{R}}^{d}}\rightarrow{\mathbb{C}} is a Lebesgue measurable function. The norm \|\cdot\|_{L^{p}} imposes certain decay conditions on the functions in L^{p} which depend on the parameter p . Using different such conditions with respect to different variables, we arrive at the definition of mixed norm Lebesgue spaces, which we recall below. Definition 1.1 . The mixed Lebesgue space L^{p,q}\left({\mathbb{R}}^{2d}\right) consists of all Lebesgue measurable functions {f}:{{\mathbb{R}}^{2d}}\rightarrow{\mathbb{C}} such that \|f\|_{L^{p,q}({\mathbb{R}}^{2d})}\equiv\|f\|_{L^{p,q}}\equiv\|f\|_{p,q}\equiv\|f_{2}\|_{q} is finite, where f_{2}(y)=\|f(\cdot,y)\|_{p}. Note that if p=q in the definition above, then L^{p,p}({\mathbb{R}}^{2d})=L^{p}({\mathbb{R}}^{2d}) . We recall the following definition of Young function. Definition 1.2 . A function {\Phi}:{[0,\infty)}\rightarrow{[0,\infty]} is called a Young function if (i) \Phi is convex, (ii) \Phi(0)=0 , (iii) \Phi(t)<\infty for some t>0 , (iv) \lim\limits_{t\rightarrow\infty}\Phi(t)=+\infty . A Young function does not have to be continuous by this definition. In fact, we do not exclude \Phi with infinite function values, hence the function \Phi(t)=\begin{cases}0&0\leq t<1,\\ t-1&1\leq t<2,\\ \infty&t\geq 2\end{cases} is a Young function. However, the only possible discontinuities for a Young function \Psi appear at points t_{0}>0 such that \Psi(t)=\infty for all t>t_{0} , since otherwise \Psi would not be convex. Evidently, this means that \Psi can have at most one jump. The reason we require (iii) in the definition above is so that each Young function has a well-defined complementary Young function, given by \Psi(t)=\sup_{s\geq 0}\left(st-\Phi(s)\right). Using the notations of [ 3 , Section 1] , we recall the definition of Lebesgue exponent