Logarithmic double phase embeddings with variable exponents: Necessary and Sufficient Conditions
Ankur Pandey and Nijjwal Karak · 2025
Read Logarithmic double phase embeddings with variable exponents: Necessary and Sufficient Conditions by Ankur Pandey and Nijjwal Karak (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-2511-13286.txt.
Introduction
Ankur Pandey and Nijjwal Karak find exact conditions for a Sobolev-type embedding in spaces with variable exponents and a logarithmic term. Their function blends two power terms with a log factor, each shaped by position in the domain. Embeddings of this kind say when control of a function and its gradient yields control of the function alone.
On bounded John domains, they prove the embedding holds in the subcritical range under mild rules on the exponents and weight. In reverse, if the embedding holds on any domain, that domain must meet a log-measure density condition. Together the results mark the sharp border for this class.
Catalog record
| CREATOR | Ankur Pandey and Nijjwal Karak |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-13286
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Logarithmic double phase embeddings with variable exponents: Necessary and Sufficient Conditions Thanks: The first author gratefully acknowledges the Council of Scientific and Industrial Research (CSIR) for awarding a Junior Research Fellowship (File Number: 09/1026(17526)/2024-EMR-I) Ankur Pandey and Nijjwal Karak Abstract . In this paper, we study the necessary and sufficient conditions in the domain for Sobolev-type embedding of the space W 1 , Φ ( ⋅ , ⋅ ) ( Ω ) W^{1,\Phi(\cdot,\cdot)}(\Omega) where Φ ( x , t ) := t p ( x ) + a ( x ) t q ( x ) log r ( x ) ( e + t ) \Phi(x,t):=t^{p(x)}+a(x)t^{q(x)}\log^{r(x)}(e+t) with 1 ≤ p ( x ) ≤ q ( x ) . 1\leq p(x)\leq q(x). We have established subcritical embedding in bounded John domains under some regularity assumptions on exponents p , p, q , q, r r , and a a . Conversely, we have proved that if the embedding holds in any domain Ω \Omega in ℝ n , \mathbb{R}^{n}, then Ω \Omega must satisfy the log-measure density condition. Keywords: Musielak-Orlicz-Sobolev spaces, Variable exponent Sobolev spaces, Logarithmic Double-Phase functions, Sobolev-type embeddings. 2020 Mathematics Subject Classification: 46E35, 46E30. 1. Introduction We assume throughout the paper that \Omega is an open subset of \mathbb{R}^{n} and we take the logarithmic variable exponent double phase function \Phi(x,t)=t^{p(x)}+a(x)t^{q(x)}\ln^{r(x)}(e+t) where a(x)\geq 0 and the variable exponents p, q and r are continuous functions defined on \Omega or \mathbb{R}^{n} , satisfying (p1) 1\leq p^{-}:=inf_{x\in\Omega}p(x)\leq sup_{x\in\Omega}p(x)=:p^{+}<\infty , (q1) 1\leq q^{-}:=inf_{x\in\Omega}q(x)\leq sup_{x\in\Omega}q(x)=:q^{+}<\infty , (r1) -\infty<r^{-}:=inf_{x\in\Omega}r(x)\leq sup_{x\in\Omega}r(x)=:r^{+}<\infty . The following three conditions in p,q, and r will also be used; the first two in the literature are known as log-Hölder continuous, and the third one is known as log-log-Hölder continuous: (p2) |p(x)-p(y)|\leq\frac{C}{\log(e+1/|x-y|)} whenever x\in\Omega and y\in\Omega (q2) |q(x)-q(y)|\leq\frac{C}{\log(e+1/|x-y|)} whenever x\in\Omega and y\in\Omega (r2) |r(x)-r(y)|\leq\frac{C}{\log(e+\log(e+1/|x-y|))} whenever x\in\Omega and y\in\Omega . For the variable exponent Sobolev space W^{1,p(\cdot)}(\Omega), the continuous Sobolev-type embedding W^{1,p(\cdot)}\hookrightarrow L^{p^{*}(\cdot)}(\Omega), when p^{+}<n, have been established in [ 9 ] for bounded domains with locally Lipschitz boundary, with the condition (p2) on the exponent p. Other Sobolev-type embeddings were considered later in [ 10 , 11 , 12 , 18 ] . Embeddings of Musielak-Orlicz-Sobolev spaces for functions \Phi of general type have been considered by several authors, [ 8 , 13 , 19 , 20 , 27 , 28 ] . Recently, Cianchi and Diening [ 6 ] have found optimal embeddings even for more general functions \Phi . The Sobolev-type embeddings for the variable exponent double phase function \Phi_{1}(x,t)=t^{p(x)}+a(x)t^{q(x)} have been shown in [ 23 ] , [ 7 ] and for two variable functions \Phi_{2}(x,t)=t^{q(x)}{(\log(e+t))}^{r(x)} in [ 21 ] . In [ 5 ] , the authors have considered the logarithmic variable exponent double phase function \Phi_{3}(x,t)=t^{p(x)}+a(x)t^{q(x)}\log(e+t) and established several properties of the space W^{1,\Phi(\cdot,\cdot)}. Here, we focus on the Sobolev-type embedding of W^{1,\Phi(\cdot,\cdot)} for a more general class of functions \Phi(x,t):=t^{p(x)}+a(x)t^{q(x)}{(\log(e+t))}^{r(x)} . Definition 1.1 . A bounded domain \Omega\subset\mathbb{R}^{n} is called a \delta -John domain, \delta>0, if there exists x_{0}\in\Omega such that each point in \Omega can be joined to x_{0} by a rectifiable path \gamma parameterized by its arc length such that B(\gamma(t),\frac{t}{\delta})\subset\Omega for all t\in[0,l(\gamma)], where l(\gamma)\leq\delta is the length of \gamma. Here we establish the embedding for W^{1,\Phi(\cdot,\cdot)}(\Omega) for a bounded John domain \Omega . Theorem 1.2 . Let \Omega be a bounded John domain. Suppose that \Phi(x,t):=t^{p(x)}+a(x)t^{q(x)}(\log(e+t))^{r(x)} and \Psi(x,t):=t^{p^{*}(x)}+(a(x))^{\frac{q^{*}(x)}{q(x)}}t^{q^{*}(x)}{(\log(e+\frac{t}{a(x)^{(q(x)-1)/q(x)}}))}^{r(x)q^{*}(x)/q(x)} with 1/p^{*}(x)=1/p(x)-1/n and 1/q^{*}(x)=1/q(x)-1/n . Assume that one of the following holds: (i) The functions p,q:\bar{\Omega}\rightarrow[1,\infty) and a:\bar{\Omega}\rightarrow[0,\infty) are Hölder continuous functions with p(x)\leq q(x) for all x\in\bar{\Omega}, \big(\frac{q}{p}\big)^{+}<1+\frac{\gamma}{n}, where \gamma is the Hölder exponent of a(x), and the function r:\bar{\Omega}\rightarrow[0,\infty) is log-log-Hölder continuous with p^{+}+r^{+}<n . (ii) The functions p:\bar{\Omega}\rightarrow[1,\infty), q:\bar{\Omega}\rightarrow[1,\infty) are log-Hölder continuous with p(x)\leq q(x) for all x\in\bar{\Omega}, a:\bar{\Omega}\rightarrow[0,\infty) is q^{+} -Hölder continuous and r:\bar{\Omega}\rightarrow[0,\infty) is any function with p^{+}+r^{+}<n. Then we have W^{1,\Phi(\cdot,\cdot)}(\Omega)\hookrightarrow L^{\Psi(\cdot,\cdot)}(\Omega). For the necessary part, it was shown in [ 14 , 15 ] that \Omega must satisfy the measure density condition to have embedding W^{1,p(\cdot)}(\Omega)\hookrightarrow L^{p^{*}}(\Omega), if p satisfies the log-Hölder condition. Note that \Omega satisfies the measure density condition if there exists a constant c>0 such that for every x in \bar{\Omega} and each R in ]0,1/2], one has |B_{R}(x)\cap\Omega|\geq cR^{n}. This condition first appeared as a necessary condition for the Sobolev embedding in [ 17 ] and later appeared as the same for other Sobolev-type embeddings as well, [ 1 , 2 , 3 , 4 , 16 , 24 , 25 , 26 ] . Recently, a weaker version of the measure density condition, namely log-measure density condition (see Definition 2.3 ), has appeared in [ 22 ] as a necessary condition for certain Orlicz-Sobolev embedding and also in [ 15 ] as a necessary condition of the Sobolev-type embedding of W^{1,p(\cdot)}(\Omega) if p is log-log-Hölder continuous in \Omega. Here we prove that if the embedding holds for W^{1,\Phi(\cdot,\cdot)} , where \Phi(x,t):=t^{p(x)}+a(x)t^{p(x)}(\log(e+t))^{r(x)} then \Omega satisfies the measure density or log-measure density conditions depending on the exponent r(\cdot). Theorem 1.3 . Let \Omega\subset\mathbb{R}^{n} . Let p:\bar{\Omega}\rightarrow[1,\infty) be log-Hölder continuous with p^{+}<n and 0\leq a(\cdot)\in L^{\infty}(\Omega). Suppose that W^{1,\Phi(\cdot,\cdot)}(\Omega)\hookrightarrow L^{\Psi(\cdot,\cdot)}(\Omega). where \Phi(x,t):=t^{p(x)}+a(x)t^{p(x)}(\log(e+t))^{r(x)} and \Psi(x,t):=t^{p^{*}(x)}+(a(x))^{\frac{p^{*}(x)}{p(x)}}t^{p^{*}(x)}{(\log(e+\frac{t}{a(x)^{(p(x)-1)/p(x)}}))}^{r(x)p^{*}(x)/p(x)} with 1/p^{*}(x)=1/p(x)-1/n and 1/q^{*}(x)=1/q(x)-1/n respectively. Then \Omega satisfies the log-measure density condition when r(x)\geq 0 for all x\in\Omega or r(x)<0 for some x\in\Omega and r(x)\geq 0 for some x\in\Omega and \Omega satisfies the measure density condition when r(x)\leq 0 for all x\in\Omega. 2. Notations ans Preliminaries Definition 2.1 . A function f:(0,\infty)\rightarrow\mathbb{R} is called almost increasing if there exists a constant c_{1}\geq 1 such that f(s)\leq c_{1}f(t) for all 0<s<t . Similarly, it is called almost decreasing if there exists a constant c_{2}\geq 1 such that c_{2}f(s)\geq f(t) for all 0<s<t. Definition 2.2 . Let f:\Omega\times(0,\infty)\rightarrow\mathbb{R} and p,q>0 .We say that f satisfies (i) (Inc)_{p}, if \frac{f(x,t)}{t^{p}} is increasing in the variable t ; (ii) (aInc)_{p}, if \frac{f(x,t)}{t^{p}} is almost increasing in the variable t ; (iii) (Dec)_{q}, if \frac{f(x,t)}{t^{q}} is decreasing in the variable t ; (iv) (aDec)_{q}, if \frac{f(x,t)}{t^{q}} is almost decreasing in the variable t . A weaker version of the measure density condition (defined in Section 1) is defined as follows. Definition 2.3 . A set \Omega\subset\mathbb{R}^{n} satisfies the log s -measure density condition if there exist two constants c>0 and \alpha>0 such that for every x in \bar{\Omega} and each R\in(0,1/2] we have cR^{s}(\log(\frac{1}{R}))^{-\alpha}\leq|B_{R}(x)\cap\Omega|. If s=n , we say that \Omega satisfies the log-measure density condition. Definition 2.4 . We say that the exponent p(\cdot) satisfies Nekvinda’s decay condition, if there exist c\in(0,1) and p_{\infty}\in[1,\infty] such that \int_{\{p(x)\not=p_{\infty}\}}c^{\frac{1}{|\frac{1}{p_{\infty}}-\frac{1}{p(x)}|}}dx<\infty. Let us recall the definition of generalized weak \Phi -functions and related standard notation. Definition 2.5 . Let (\Omega,\Sigma,\mu) be a complete \sigma -finite measurable space. A function \Phi:\Omega\times[0,\infty)\rightarrow[0,\infty] is said to be a generalized \Phi -prefunction if \Phi is measurable in the first variable, increasing in the second variable and satisfies \Phi(x,0)=0 , \lim_{t\rightarrow 0^{+}}\Phi(x,t)=0 , and \lim_{t\rightarrow+\infty}\Phi(x,t)=+\infty for a.e. x\in\Omega. We say that \Phi is a generalized weak \Phi -function if it satisfies (aInc)_{1} on \Omega\times[0,\infty). We denote the set of generalized weak \Phi -functions by \Phi_{w}(\Omega,\mu) . The generalized weak \Phi