Abstract
Our aim is to characterize the homogeneous fractional Sobolev–Slobodeckiĭ spaces \(\mathcal {D}^{s,p} (\mathbb {R}^n)\) and their embeddings, for \(s \in (0,1]\) and \(p\ge 1\). They are defined as the completion of the set of smooth and compactly supported test functions with respect to the Gagliardo–Slobodeckiĭ seminorms. For \(s\,p < n\) or \(s = p = n = 1\) we show that \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) is isomorphic to a suitable function space, whereas for \(s\,p \ge n\) it is isomorphic to a space of equivalence classes of functions, differing by an additive constant. As one of our main tools, we present a Morrey–Campanato inequality where the Gagliardo–Slobodeckiĭ seminorm controls from above a suitable Campanato seminorm.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
1.1 Fractional Sobolev spaces
The aim of this paper is to shed light on an important topic in the theory of fractional Sobolev spaces. This family of spaces is conveniently presented e.g. in [1, 17, 19, 29]. It is common to define the fractional Sobolev spaces \(W^{s,p} (\mathbb {R}^n)\) in the Sobolev–Slobodeckiĭ form. Thus, for \(s \in (0,1)\) and \(1\le p<+\infty \) we define the normalized Gagliardo–Slobodeckiĭ seminorm by
Then,
is a Banach space endowed with the non-homogeneous norm
A word about the convenience of the extra factor \(s\,(1-s)\) is in order. Indeed, \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\) can be thought as a real interpolation quantity, with parameter s, between the two quantities
see for example [6] or [9]. It is then natural to expect the following asymptotic behaviour
and
see [20] for the first result and [4] for the second one. For this reason, the factor \(s\,(1-s)\) is incorporated in the definition of the seminorm and the limit cases \(s=0\) and \(s=1\) are defined accordingly by
There are well-known embeddings of these spaces \(W^{s,p} (\mathbb {R}^n)\) into \(L^q(\mathbb {R}^n)\) for suitable \(q\ge 1\) for which we refer to the classical monographs like [1, 17].
The particular case \(s=(p-1)/p\) has a peculiar theoretical importance, since in this case
can be identified with the trace space of functions in \(W^{1,p}(\mathbb {H}^{n+1}_+)\), where \(\mathbb {H}^{n+1}_+=\mathbb {R}^n\times [0,+\infty )\). More generally, it can be proved that \(W^{s,p}(\mathbb {R}^n)\) coincides with the trace space of the weighted Sobolev space \(\mathcal {W}_s^{1,p}(\mathbb {H}^{n+1}_+)\), defined as
where we have used the notation \((x,y)\in \mathbb {H}^{n+1}_+\), with \(x\in \mathbb {R}^n\) and \(y\in [0,+\infty )\). See [18, Section 5] for more details. The reader may also consult the recent paper [22], containing some generalizations.
1.2 Motivation for the homogeneous Sobolev spaces
Before we present the main results of this paper, we discuss some motivations for the study of a particular class of fractional Sobolev spaces. Recently, there has been a surge of interest towards the study of nonlocal elliptic operators, that arise as first variations of Gagliardo–Slobodeckiĭ seminorms. The leading example is given by the fractional Laplacian of order s of a function u, indicated by the symbol \((-\Delta )^s u\), which in weak form reads as
for all \(\varphi \in C^\infty _c(\mathbb {R}^n)\), up to a possible normalization factor. Observe that this is nothing but the first variation of the functional
up to the factor \(s\,(1-s)\). More generally, one could take a general exponent \(1<p<+\infty \) and obtain accordingly the fractional p-Laplacian of order s of a function u, which we denote \((-\Delta _p)^s u\), and is defined in weak form by
for all \(\varphi \in C^\infty _c(\mathbb {R}^n)\), up to a multiplicative factor. This is a nonlocal and nonlinear operator which has been extensively studied in recent years, see [31, 32] and references therein. In order to motivate the studies performed in this paper, let us consider the quasilinear nonlocal elliptic problem
under suitable assumptions on the source term f. In order to prove existence of a weak solution, it would be natural to use the Direct Method in the Calculus of Variations. This would lead to the problem of minimizing the energy functional
which is naturally associated to our equation. However, it is not clear the functional space where this minimization problem should be posed. For example, one could try to pose the problem in the previously introduced space \(W^{s,p}(\mathbb {R}^n)\), but it is easily seen that this does not fit at all. Indeed, the functional (1.2) is not weakly coercive on this space, unless we are in the trivial situation \(f\equiv 0\). Of course, the problem is that the functional (1.2) can not permit to infer any control on the \(L^p\) norm of minimizing sequences. This in turn is related to the fact that the Poincaré inequality
fails to be true on the whole \(\mathbb {R}^n\) for any \(c>0\). This can be easily seen by using the invariance of \(\mathbb {R}^n\) with respect to scalings \(x\mapsto \lambda \,x\) and a simple dimensional analysis of the two norms.
It turns out that the natural spaces to work with are the homogeneous Sobolev spaces \(\mathcal {D}^{s,p} (\mathbb {R}^n)\). They are defined by
We recall that \(W^{s,p}(\mathbb {R}^n)\) and \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) are particular instances of the huge family of Besov spaces. In this respect, we stress that a mention of homogeneous Besov spaces can be found for example in [1, Remark 7.68], [3, Chapter 6, Section 3], [19, Chapter 10, Section 1] and [30, Chapter 3, Section 4], among others.
The notation \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) adopted here is reminiscent of the historical one, introduced by Deny and Lions in their paper [10]. This reference has been among the first papers to study homogeneous spaces, obtained by completion of \(C^\infty _c(\mathbb {R}^n)\).
Another frequently encountered notation for homogeneous Sobolev spaces is \(\dot{W}^{s,p}(\mathbb {R}^n)\), see for example Petree’s paper [26]. However, usually this notation is used for spaces of functions identified modulo constants for \(s\in (0,1]\). We will adopt the same convention in this paper.
It is our aim to examining the space \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) more closely and connect it with spaces of the type \(\dot{W}^{s,p}(\mathbb {R}^n)\), as we will explain in a moment.
1.3 Completions
Before presenting the main results of the paper, let us briefly recall some basic facts about the completion process. We start from the seminorm \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)} \) for \(0<s \le 1\) and \(1<p<\infty \). It is not difficult to see that this turns out to be a norm on the space \(C_c^\infty (\mathbb {R}^n)\).
However, the normed space
is not complete. By definition, its completion is the quotient space of the set of sequences \((u_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) which are Cauchy for the norm \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\), under the expected equivalence relation
For each equivalence class \(U = \{(u_m)_{m\in \mathbb {N}}\}_{s,p}\in \mathcal {D}^{s,p}(\mathbb {R}^n)\), we define its norm in terms of a representative as
It is easily seen that such a definition is independent of the chosen representative. With this construction,
is a Banach space. Note that all functions \(u \in C_c^\infty (\mathbb {R}^n)\) can be naturally embedded into \(\mathcal {D}^{s,p} (\mathbb {R}^n)\) by the constant sequence \(u_m = u\). By definition, this representation of \(C_c^\infty (\mathbb {R}^n)\) is dense in \(\mathcal {D}^{s,p} (\mathbb {R}^n)\).
In the case \(s=1\), it is well-known that \(\mathcal {D}^{1,p}(\mathbb {R}^n)\) is a subspace of the space of distributions \(\mathcal {D}'(\mathbb {R}^n)\) if and only if \(1\le p<n\). In this case, this can be identified with the functional space
thanks to the celebrated Sobolev inequality. The case \(p=2\) is contained in Deny and Lions, [10, Théorème 4.4 and Remark 4.1]. The general case can be found for example in [19, Chapter 15]. On the contrary, the case \(p\ge n\) is much more delicate, since in this case \(\mathcal {D}^{1,p}(\mathbb {R}^n)\) is not even a subspace of \(\mathcal {D}'(\mathbb {R}^n)\). A concrete characterization of \(\mathcal {D}^{1,p}(\mathbb {R}^n)\) as a space of equivalence classes of functions modulo constants seems to belong to the folklore on the subject, though we have not been able to find a proper reference in the literature. Our presentation will cover this case, as well.
1.4 Main results: the three ranges
The main question we address in these notes is the characterization of \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) and the study of some of its embeddings into suitable sets of functions. This will be possible for some exponents, while for other exponents the embedding occurs into a space of equivalence classes of functions modulo constants, as we are now going to explain.
More precisely, in order to answer to these questions, we will need to distinguish three cases, according to the different behaviours of
with respect to scalings of the form \(x\mapsto \lambda \,x\), with \(\lambda >0\). By a simple change in variable, we have that
This shows that the relation between \(s\,p\) and n provides significantly different results. Consequently, there are three different situations:
1.4.1 Subconformal case \(s\,p<n\)
The natural inclusion of \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) into a Lebesgue space of functions succeeds. More precisely, the completion can be identified with a functional space, i. e., we have
see Theorem 3.1 below. The main tool here is the fractional Sobolev inequality. This mimics in some sense what happens for \(W^{s,p} (\mathbb {R}^n)\), that can be characterized by (1.1) as the completion of \(C_c^\infty (\mathbb {R}^n)\) with respect to the \(W^{s,p} (\mathbb {R}^n)\) norm.
1.4.2 Superconformal case \(s\,p > n\)
Here, the Sobolev inequality is not available and we have to replace it by Morrey’s inequality, see Eq. (2.11) below. However, unlike the case \(s\,p<n\), the elements in \(\mathcal {D}^{s,p} (\mathbb {R}^n)\) can not be uniquely represented by functions. Indeed, when \(s\,p > n\), there exist sequences \((\varphi _m)_{m\in \mathbb {N}}\) such that
as m goes to \(\infty \). These sequences are known as null-sequences. Hence, any sequence \((u_m)_{m\in \mathbb {N}}\subset C_c^\infty (\mathbb {R}^n)\) which is Cauchy in the norm \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\) is equivalent to the sequence
for any constant \(C\in \mathbb {R}\). Observe that this implies in particular that now all constant functions are equivalent to the null one in \(\mathcal {D}^{s,p} (\mathbb {R}^n)\).
Furthermore, one can show that functions that are approximated by equivalent Cauchy sequences actually coincide up to a constant. This allows to show that \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) can be identified with a space of equivalence classes of Hölder continuous functions differing by an additive constant, i. e. we have
where for \(0<\alpha \le 1\)
and \(\sim _C\) is the equivalence relation
We refer to Theorem 4.4 below, for complete details.
1.4.3 Conformal case \(s\,p = n\)
This is the most delicate case. Whenever \(s<n\) (i. e., unless \(s=n=1\)), it is still possible to prove existence of a sequence \((\varphi _m)_{m\in \mathbb {N}}\) such that properties (1.4) hold. However, the construction of such a sequence is now more involved. Since the seminorm is scale invariant in this case, such a construction can not be just based on scalings. As in the local case \(s=1\), one has to consider a suitable sequence of truncated and rescaled logarithms (see Lemma 5.1 below). This is reminiscent of the optimal sequence for the Moser-Trudinger inequality, see for example [25, Section 5] for the fractional case.
This permits to show that also in the case \(s\,p=n\) (provided \(s < n\)), we can approximate the constant functions by functions in the null class. Hence, this case behaves like \(s\,p > n\) and \(\mathcal {D}^{s,n/s}(\mathbb {R}^n)\) can be identified with a space of equivalence classes of BMO functions differing by an additive constant, i.e. we have
see Theorem 5.3 below.
Still, there will be room for a small surprise. Indeed, we will show that the limiting case \(s=p=n=1\), which still falls in the conformal regime, behaves like \(s\,p < n\). In other words, the homogeneous Sobolev space \(\mathcal {D}^{1,1}(\mathbb {R})\) is actually a function space and we have
where \(C_0(\mathbb {R})\) is the space of continuous functions vanishing at infinity, see Theorem 5.5 below. In this way we complete the characterization of the spaces.
Remark 1.1
After completing this work, we became aware of the interesting recent paper [23], dealing with the same issue here addressed, but for a different scale of fractional Sobolev spaces. Namely, the authors of [23] deal with the so-called Bessel potential spaces (sometimes also called Liouville spaces), which are defined in terms of the Fourier transform. In [23, Theorem 2], they give a concrete realization of the homogeneous version of these spaces.
Our results only partially superpose with those of [23] and, in any case, the proofs are different.
Indeed, the results of [23] are based on Harmonic Analysis techniques and contain ours only for the case \(p=2\) and \(0<s<1\), and for \(1<p<\infty \) and \(s=1\).
1.5 Plan of the paper
We start with Sect. 2, where some basic facts about BMO and Campanato spaces are recalled. These tools are particularly useful to handle the cases \(s\,p\ge n\). In this part, an important result is Theorem 2.4, which relates the Gagliardo–Slobodeckiĭ seminorm and a Campanato seminorm. This yields as corollaries several important inequalities: a fractional Poincaré–Wirtinger inequality (see Corollary 2.5) and the fractional Morrey inequality for \(s\, p > n\) (see Corollary 2.7).
We devote Sects. 3, 4 and 5 to prove the characterisation of \({\mathcal {D}}^{s,p}\) in the cases \(s\,p<n\), \(s\,p>n\), and \(s\,p=n\), respectively. We introduce in each case suitable structural lemmas.
We conclude the paper with two appendices on approximation lemmas, which will allow to show that the elements in \(\dot{W}^{s,p} (\mathbb R^n)\) can be approximated by functions in \(C_c^\infty (\mathbb {R}^n)\). The aim is to approximate a function u in the Gagliardo–Slobodeckiĭ seminorm by sequences of the type \((u *\rho _m)\, \eta _m\), where \(\rho _m\) are standard mollifiers and \(\eta _m\) are cut-off functions. In Appendix A we study the convolution, while in Appendix B we prove several truncation lemmas, which allow to estimate the effect of multiplying by cut-off functions.
2 Preliminaries
2.1 BMO and Campanato spaces
At first, we need to recall definitions and some basic facts about bounded mean oscillation functions and Campanato spaces. As already announced, this will be particularly useful to deal with the cases \(s\,p>n\) and \(s\,p=n\). We will indicate by \(B_r(x_0)\) the n-dimensional open ball with center \(x_0\in \mathbb {R}^n\) and radius \(r>0\). The symbol \(\omega _n\) will stand for the measure of \(B_1(0)\).
A function of bounded mean oscillation is a locally integrable function u such that the supremum of its mean oscillations is finite. More precisely, for every \(u\in L^1_\mathrm{loc}(\mathbb {R}^n)\) we define
where
Then we define the space of functions with bounded mean oscillation as
This space was introduced by John and Nirenberg in [16]. The BMO space is a borderline space, which plays a key role in different areas of Mathematical Analysis, as a natural replacement of \(L^\infty (\mathbb {R}^n)\) in a large number of results, for instance in interpolation. Fefferman and Stein characterized this space as the dual of the Hardy space \(\mathcal {H}^1\), see [11, Theorem 1] and [12, Theorem 2]. Another important appearance of this space is in Elliptic Regularity Theory: indeed, the logarithm of a positive local solution to an elliptic partial differential equation is a locally BMO function. This observation is a crucial step in the classical proof by Moser of Harnack’s inequality, see [24].
It turns out that the BMO space can be seen as a particular instance of the larger family of Campanato spaces, see [7]. For \(1\le p<+\infty \) and \(0\le \lambda \le n+p\), for every \(u\in L^p_\mathrm{loc}(\mathbb {R}^n)\) we define the seminorm
Accordingly, we introduce the Campanato space
Notice that since \(|B_\varrho (x_0)| = \omega _n\, \varrho ^n\) we have
We recall that for \(n<\lambda \le n+p\), we have
see [14, Chapter 2, Section 3]. The constant \(C=C(\lambda ,n,p)>0\) blows-up as \(\lambda \searrow n\).
We must point out that both \([\,\cdot \,]_{BMO(\mathbb {R}^n)}\) and \([\,\cdot \,]_{\mathcal {L}^{p,\lambda }(\mathbb {R}^n)}\) are only seminorms on their relevant spaces, as they do not detect constants, i.e. the seminorm of constant functions is zero. In order to get a normed space (actually, a Banach space), we need to consider their homogeneous version defined as quotient spaces
with the equivalence relation
We will denote by \(\{ u\}_{C}\) the class of u with respect to this relation.
An interesting result, which can be found for example in [14, Proposition 2.5 and Corollary 2.3], says that when \(\lambda = n\) all the Campanato spaces are isomorphic and we have
On the contrary, for \(\lambda \not =n\) the spaces \(\mathcal L^{p,\lambda }(\mathbb {R}^n)\) and \(\mathcal {L}^{q,\lambda }(\mathbb {R}^n)\) do not coincide, for \(p\not =q\).
Remark 2.1
As pointed out in [27, Chapter 4, §1.1.1], the definition of the BMO space can be equivalently given by taking hypercubes instead of balls. The same remark applies to Campanato spaces.
2.2 Weighted integrability for some Campanato spaces
Functions belonging to \(\mathcal {L}^{p,n}(\mathbb {R}^n)\) enjoy a suitable weighted global integrability condition. More precisely, we have the following
Lemma 2.2
Let \(1\le p<+\infty \) and \(R>0\). For every \(u\in \mathcal {L}^{p,n}(\mathbb {R}^n)\) such that
we have
for some constant \(C=C(n,p)>0\).
Proof
The proof is an adaptation of that of [12, Equation (1.2)]. The final outcome is slightly better. We observe that it is sufficient to prove (2.3) for \(R=1\). The general case then follows by a standard scaling argument.
We first fix some shortcut notation, for the sake of simplicity. For every \(k\in \mathbb {N}\), we set
and observe that \(u_0=0\), by assumption. We start by estimating the difference \(|u_{k+1}-u_{k}|\). We have
By observing that
we can divide both sides by \(|B_{2^{k}}(0)|\) and get
By using this estimate and the triangle inequality, we then get for every \(j\in \mathbb {N}{\setminus }\{0\}\)
We now get
In particular, this implies for every \(j\in \mathbb {N}{\setminus }\{0\}\)
We now observe that for every \(x\in B_{2^j}(0){\setminus } B_{2^{j-1}}(0)\)
If we use this estimate in the previous inequality, we get
We further divide both sides by \(j^2\), so to get
On the left-hand side, we use that for every \(x\in B_{2^j}(0){\setminus } B_{2^{j-1}}(0)\) it holds
This in turn implies that
possibly for a different constant \(C>0\). If we now sum over \(j\ge 1\) we get
We are only left with observing that we have (recall that u has average 0 in \(B_1(0)\))
and
Thus we get
as well. By summing up (2.4) and (2.5), we get (2.3) for \(R=1\), as desired. \(\square \)
Remark 2.3
Due to Eq. (2.2), the previous weighted estimate applies to BMO, as well.
2.3 A Morrey–Campanato—type inequality and applications
We now prove an inequality relating the Gagliardo–Slobodeckiĭ and Campanato seminorms. This will give us, as corollaries, fractional versions of the Poincaré–Wirtinger and Morrey inequalities.
Theorem 2.4
Let \(s\in (0,1)\) and \(1\le p<+\infty \). Then for every \(u\in C^\infty _c(\mathbb {R}^n)\) we have
for a constant \(C=C(n,p)>0\).
Proof
We fix \(x_0\in \mathbb {R}^n\) and \(\varrho >0\), then for every \(u,v\in C^\infty _c(\mathbb {R}^n)\) by Minkowski inequality we get
In the second estimate, we used Jensen’s inequality. We now apply the standard Poincaré–Wirtinger inequality (see for example [14, Theorem 3.17]) in order to control the last term
for a constant \(C=C(n,p)>0\). Then the last two displays imply the estimate
This estimate is valid for every \(v\in C^\infty _c(\mathbb {R}^n)\). Thus, if we define the \(K-\)functional
by taking the infimum over v in the last estimate, we get
By raising to power p and integrating over \((0,+\infty )\) with respect to the singular measure \(d\varrho /\varrho \), we get
From [6, Proposition 4.5], we have that
for a constant \(C=C(n,p)>0\). Up to now, we obtained
We now use that
thus we get for \(r>0\)
Since \(r>0\) and \(x_0\in \mathbb {R}^n\) are arbitrary, from (2.6) we thus obtain that
By recalling that the quantity in the left-hand side is equivalent to the Campanato seminorm \(\mathcal {L}^{p,sp}\) (see [14, Remark 2.2]), we get
for some \(C=C(n,p)>0\). In a similar way, we observe that for \(r>0\)
As before, from (2.6) we get
and thus
for some \(C=C(n,p)>0\). If we now multiply (2.7) by \((1-s)\), (2.8) by s and then take the sum, we get the claimed estimate. \(\square \)
As a first straightforward consequence of Theorem 2.4, we get the following
Corollary 2.5
(Fractional Poincaré–Wirtinger inequality) Let \(s\in (0,1)\) and \(1\le p<+\infty \). Then for every \(u\in C^\infty _c(\mathbb {R}^n)\) and every \(x_0\in \mathbb {R}^n\), \(R>0\), we have
for a constant \(C=C(n,p)>0\).
Remark 2.6
As a simple consequence of the previous result, we also have the following more flexible inequality: for every \(u\in C^\infty _c(\mathbb {R}^n)\) and every \(x_0\in \mathbb {R}^n\), \(0<r\le R\), we have
for a possibly different constant \(C=C(n,p)>0\). Indeed, it is sufficient to observe that, due to Jensen’s inequality,
An application of Corollary 2.5 now leads to the claimed estimate (2.10).
Theorem 2.4, together with the estimate (2.1), also implies the following result. We include in the statement the case \(s=1\), which is classical, see for example [14, Theorem 3.9].
Corollary 2.7
(Fractional Morrey’s inequality) Let \(s\in (0,1]\) and \(1\le p<+\infty \) be such that \(s\,p> n\). Then for every \(u\in C^\infty _c(\mathbb {R}^n)\) we have
for a constant \(C=C(n,p,s)>0\). Such a constant may be taken independent of s, whenever \(s-n/p\ge \delta _0\), for some \(\delta _0>0\). In this case, it has the form \(C=C(n,p,\delta _0)>0\).
Remark 2.8
The previous result is well-known, see for example [26, Théorème 8.2] for a proof using a different interpolation-type argument. The main focus here is on the presence of the scaling factor \(s\,(1-s)\), which is incorporated in our definition of the Gagliardo–Slobodeckiĭ seminorm. If one is not interested in keeping track of this factor, actually the proof simplifies, see for example [13, Lemma 2.3].
We conclude this section with a variant of the Poincaré–Wirtinger inequality. We do not pay too much attention to the quality of the constant: the resulting outcome will be sufficient for our purposes.
Lemma 2.9
Let \(s\in (0,1)\) and \(1\le p<+\infty \). Let \(u\in L^1_\mathrm{loc}(\mathbb {R}^n)\) be such that
The for every \(0<r<R\) and \(x_0\in \mathbb {R}^n\), we have
for a constant \(C=C(n,p,s)>0\).
Proof
The proof is quite straightforward, it is the same that can be found in [21, page 297], for example. By using Jensen’s inequality, we have
We now observe that
By using this simple fact in the previous estimate, we get the desired conclusion. \(\square \)
3 Characterisation for \(s\,p < n\)
This case is similar to the local case (i. e. \(s=1\)) for \(p<n\). Indeed, in this range we obtain an embedding into a Lebesgue space, by means of the fractional Sobolev inequality
for some constant \(\mathcal {S}_{s,p}>0\). Here \(p^\star _s\) is the critical Sobolev exponent, defined by
We refer to [19, Theorem 10.2.1] for an elementary proof of (3.1). See also [26, Théorème 8.1] for an older proof, based on real interpolation techniques.
By using inequality (3.1), it is possible to give a concrete characterization of the completion \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) as a functional space.
Theorem 3.1
Let \(s \in (0,1]\) and \(1\le p<+\infty \) be such that \(s\,p < n\). We indicate by \(\dot{W}^{s,p}(\mathbb {R}^n)\) the space
We endow this space with the norm
Then this is a Banach space, having \(C^\infty _c(\mathbb {R}^n)\) as a dense subspace. Moreover, there exists a linear isometric isomorphism
In other words, the space \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) can be identified with \(\dot{W}^{s,p}(\mathbb {R}^n)\).
Proof
It is easy to see that \(\dot{W}^{s,p}(\mathbb {R}^n)\) is a normed vector space. The fact that this is a Banach space will follow from the claimed isometry, that we are going to construct at the end of the proof.
We now divide the rest of the proof in three parts.
Part 1: density of smooth functions. We prove that \(C^\infty _c(\mathbb {R}^n)\) is dense in \(\dot{W}^{s,p}(\mathbb {R}^n)\). We need to prove that for every \( u \in \dot{W}^{s,p} (\mathbb {R}^n)\), there exists a sequence \((u_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\) such that
In order to construct the sequence \((u_m)_{m\in \mathbb {N}}\), we consider the sequence of smoothing kernels \((\rho _m)_{m\ge 1}\) as in the statement of Lemma A.1. Moreover, we introduce a sequence of cut-off functions \( \eta _j\in C^\infty _c ({\mathbb {R}}^n)\) with \(\mathrm {supp\,} \eta _j \subset B_{2j}\) such that
By Lemma B.1, for every \(m\ge 1\) we have
Thus, for every \(m\ge 1\) we can choose \(j_m\in \mathbb {N}\) such that
We finally set
then this sequence has the desired approximation property. Indeed, observe that by the triangle inequality we have
By taking the limit as m goes to \(\infty \) and appealing to Lemma A.1, we get the conclusion.
Part 2: Cauchy sequences in \(\mathcal {D}^{s,p}(\mathbb {R}^n)\). We take a sequence \((u_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\), which is a Cauchy sequence with respect to the norm
By using the fractional Sobolev inequality (3.1), we have that this is a Cauchy sequence in \(L^{p^\star _s}(\mathbb {R}^n)\), as well. The latter being a Banach space, we get that the sequence converges strongly in \(L^{p^\star _s}(\mathbb {R}^n)\) to a function \(u\in L^{p^\star _s}(\mathbb {R}^n)\). Furthermore, we can show that \(u\in \dot{W}^{s,p}(\mathbb {R}^n)\).
Indeed, if we fix \(\varepsilon >0\), then by definition of Cauchy sequence there exists \(n_\varepsilon \in \mathbb {N}\) such that
In particular, by Minkowski inequality we get
This shows that the sequence
is bounded in \(L^p(\mathbb {R}^n\times \mathbb {R}^n)\). By using that \(u_m\) converges almost everywhere to u (up to a subsequence), Fatou’s Lemma entails
i. e. \(u\in \dot{W}^{s,p}(\mathbb {R}^n)\).
We now observe that \(L^p(\mathbb {R}^n\times \mathbb {R}^n)\) is a Banach space, thus the Cauchy sequence (3.3) converges strongly in \(L^p(\mathbb {R}^n\times \mathbb {R}^n)\). By uniqueness, the limit must coincide with
In conclusion, we obtain that the Cauchy sequence \((u_m)_{m\in \mathbb {N}}\) converges with respect to the Gagliardo–Slobodeckiĭ seminorm to an element of \(\dot{W}^{s,p}(\mathbb {R}^n)\).
Part 3: construction of the isometry. We now take \(U\in \mathcal {D}^{s,p}(\mathbb {R}^n)\) and choose a representative of this equivalence class, i.e. \(U=\{(u_m)_{m\in \mathbb {N}}\}_{s,p}\). Thanks to Part 2, we know that \((u_m)_{m\in \mathbb {N}}\) converges to a function \(u\in \dot{W}^{s,p}(\mathbb {R}^n)\). We then define
Observe that this is well-defined, since for any other representative \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\) belonging to the class U, we still have
Moreover, it is easy to see that \(\mathcal {J}\) is linear. It is also immediate to obtain that this is an isometry, since
We are left with proving that \(\mathcal {J}\) is surjective. From Part 1 we know that for every \(u\in \dot{W}^{s,p}(\mathbb {R}^n)\) there exists a sequence \((u_m)_{n\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\) such that
In particular, this is a Cauchy sequence with respect to the Gagliardo–Slobodeckiĭ seminorm. Thus we get
This concludes the proof. \(\square \)
Remark 3.2
We note that
with continuous inclusion. The inclusion in \(L^p(\mathbb {R}^n)\) is a straightforward consequence of the definition (1.1) of \(W^{s,p}(\mathbb {R}^n)\). On the other hand, the inclusion in \(L^{p^\star _s} (\mathbb {R}^n)\) follows from the fractional Sobolev inequality and the density of \(C^\infty _c(\mathbb {R}^n)\) functions in \(W^{s,p}(\mathbb {R}^n)\).
We can exploit this summability information to see that
For example, it is not difficult to see that the function
is such that
Indeed, \(\varphi \not \in L^p(\mathbb {R}^n)\) thanks to the choice of \(\alpha \). In order to see that \(\varphi \) has a finite Gagliardo–Slobodeckiĭ seminorm, it is useful to decompose the seminorm as follows
The integrals \(\mathcal {I}_2\) and \(\mathcal {I}_4\) are finite, thanks to the Lipschitz character of \(\varphi \). The third integral \(\mathcal {I}_3\) is finite, by using that \(\varphi \in L^\infty (\mathbb {R}^n)\) and that the function
is uniformly bounded for \(x\in B_1(0)\). Finally, for the finiteness of \(\mathcal {I}_1\), it is sufficient to observe thatFootnote 1
so that
The last double integral is then finite, by appealing to [5, Lemma A.1].
Remark 3.3
We take the occasion to recall that the interesting question of determining the sharp constant in (3.1) is still open, except for the case \(p=2\), solved in [8]. It is clear that the sharp constant is given by
The relevant Euler-Lagrange optimality condition is a nonlinear eigenvalue-type equation involving the operator \((-\Delta _p)^s\), already presented in the Introduction. Namely, an extremal for the previous problem has to be a constant sign solution of
Some properties of solutions to (3.4) have been investigated in [5, Theorem 1.1].
4 Characterisation for \(s\,p > n\)
Instead of the fractional Sobolev inequality, in this range we have the fractional Morrey inequality, see Corollary 2.7. However, unlike Sobolev’s inequality, inequality (2.11) does not detect constants. Even worse, in this range constant functions can be approximated by sequences in \(C_c^\infty (\mathbb {R}^n)\) with respect to the Gagliardo–Slobodeckiĭ seminorm.
Lemma 4.1
Let \(s \in (0,1]\) and \(s\,p > n\). There exists a sequence \((\varphi _m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) such that
as \(m\rightarrow +\infty \). Hence, \((\varphi _m)_{m\in \mathbb {N}}\) is equivalent in \(\mathcal {D}^{s,p} (\mathbb {R}^n)\) to zero, although its pointwise limit is 1.
Proof
The proof is just based on the scaling properties of the Sobolev–Slobodeckiǐ seminorm, as in the local case. Let \(\varphi \in C_c^\infty (B_2)\) be a non-negative cut-off function, such that \(\varphi \) coincides identically with 1 on \(B_1\). We define the rescaled sequence
By recalling (1.3), we have that
The conclusion now follows, thanks to the fact that \(n/p-s<0\). \(\square \)
Remark 4.2
Notice that this construction is linked with the relative \((s,p)-\)capacity of a set \(\omega \Subset \Omega \), defined as
see for example [33]. For \(s=1\) this value can be explicitly computed, and the relevant Euler-Lagrange equation in linked to the usual \(p-\)Laplacian.
We need the following technical result.
Lemma 4.3
Let \(s \in (0,1]\) and \(s\,p > n\), we define \(\alpha =s-n/p\). Let \((u_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) be a Cauchy sequence with respect to \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\). Then there exists another sequence \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\subset C_c^\infty (\mathbb {R}^n)\) such that:
-
we have
$$\begin{aligned} \lim _{m\rightarrow \infty }[u_m - {\widetilde{u}}_m]_{W^{s,p} (\mathbb {R}^n)}=0, \end{aligned}$$ -
\({\widetilde{u}}_m\) converges uniformly over compact sets to a \(\alpha -\)Hölder continuous function u.
Moreover, this function u is such that \(u(0) = 0\),
and
Proof
We construct the sequence as follows. We take
and consider \(\varphi _m\) given by Lemma 4.1. By construction \(M_m\) diverges to \(\rightarrow +\infty \) as m goes to \(+\infty \). Then we define
We now show that \(\widetilde{u}_m\) has the claimed properties. At first, by construction we have
which implies that
as desired. Observe that this implies that \((\widetilde{u}_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) is still a Cauchy sequence with respect to \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\).
In order to infer the uniform convergence, it is sufficient to show that \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\) is a Cauchy sequence in C(K), for every \(K\subset \mathbb {R}^n\) compact set. This follows directly, by applying (2.11) to \(\widetilde{u}_m - \widetilde{u}_k\), i. e.
Hence, it converges uniformly to some function \(u \in C(K)\). Since K is arbitrary and the limit is unique, u is defined for every \(x \in \mathbb {R}^n\). Moreover, it holds \(u(0)=0\), since we have \({\widetilde{u}}_m(0)=0\) by construction.
By using that \(({\widetilde{u}}_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) is still a Cauchy sequence with respect to \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\) and arguing as in Part 2 of the proof of Theorem 3.1, we deduce (4.1). The estimate (4.2) can then be obtained by passing to the limit in (2.11). \(\square \)
The major difference with respect to the case \(s\,p<n\) is that now the elements in \(\mathcal {D}^{s,p} (\mathbb {R}^n)\) can not be uniquely represented by functions. Indeed, when \(s\,p > n\), any sequence \((u_m)_{m\in \mathbb {N}}\subset C_c^\infty (\mathbb {R}^n)\) which is Cauchy in the norm \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\) is equivalent to the sequence
for any constant \(C\in \mathbb {R}\). Here \(\varphi _m\) is the same as in Lemma 4.1.
However, one can show that functions that are approximated by equivalent Cauchy sequences, actually coincide up to a constant. In other words, the homogeneous space \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) can be identified with a space of equivalence classes of functions differing by an additive constant.
More precisely, we have the following characterization, which is the main result of this section.
Theorem 4.4
Let \(s \in (0,1]\) and \(s\,p > n\), we set \(\alpha = s -n/p\). We consider the quotient space
where
We will denote by \(\{ u\}_{C}\) the equivalence class of u with respect to this relation. We endow this space with the norm
Then this is a Banach space and there exists a linear isometric isomorphism
In other words, the space \(\mathcal {D}^{s,p}(\mathbb {R}^n)\) can be identified with \(\dot{W}^{s,p}(\mathbb {R}^n)\).
Proof
It is immediate to see that \(\dot{W}^{s,p} (\mathbb {R}^n)\) is a normed vector space. Indeed, constant functions all belong to the equivalence class \(\{0\}_C\).
We now construct the isometry. The fact that \(\dot{W}^{s,p} (\mathbb {R}^n)\) is a Banach space will follow at once. For any class
we may apply Lemma 4.3 and consider the new Cauchy sequence \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\). By construction, we have
and by Lemma 4.3, we know that \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\) converges to some function
We may identify U with the equivalence class \(\{u\}_C\), i. e. we define \(\mathcal {J}(U) =\{u\}_C\).
Observe that this is well-defined, since for any other representative \((v_m)_{m\in \mathbb {N}}\) belonging to the class U, we still have
Let us show that the map \(\mathcal {J}\) is linear. Indeed for every \(U,V\in \mathcal {D}^{s,p}(\mathbb {R}^n)\) and \(\alpha ,\beta \in \mathbb {R}\), we consider \(\alpha \,U+\beta \,V\). By choosing \((u_m)_{m\in \mathbb {N}}\) a representative for U and \((v_m)_{m\in \mathbb {N}}\) a representative for V, we can apply Lemma 4.3 to both sequences and obtain two new sequences \(({\widetilde{u}}_m)_{m\in \mathbb {N}}, (\widetilde{v}_m)_{m\in \mathbb {N}}\)
and two functions
such that we have
By observing that
and using that
we get
as desired.
Moreover, by construction we have
which implies that this is an isometry.
We still have to show that \(\mathcal {J}\) is surjective. For every equivalence class \(\{v\}_C\in \dot{W}^{s,p}(\mathbb {R}^n)\), we may select the representative v in such a way that
Then we can construct a sequence \((v_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\) such that
In order to do this, we can repeat the construction of Part 1 in the proof of Theorem 3.1, up to some modifications that we are going to detail. More precisely, we introduce a sequence of cut-off functions \( \eta _j\in C^\infty _c ({\mathbb {R}}^n)\) with \(\mathrm {supp\,} \eta _j \subset B_{2j}(0)\) such that
and observe that by Lemma B.2, for every \(m\ge 1\)
for a constant \(C>0\) independent of j. This shows that the sequence
weakly converges, up to a subsequence. The weak limit is given by the null function, since \(1-\eta _j\) converges to 0, locally uniformly. In order to upgrade this convergence, we can apply Mazur’s Lemma to infer that there exists a new sequence made of convex combinations of (4.4), which converges strongly to 0 in \(L^p(\mathbb {R}^n\times \mathbb {R}^n)\).
Thanks to the form of the sequence (4.4), this finally implies that there exists \((\widetilde{\eta }_j)_{j\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\), such that
in other words, we have
Thus, for every \(m\ge 1\) we can choose \(j_m\in \mathbb {N}\) such that
If we now define the sequence \((v_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\) by
it is easy to see that this verifies (4.3), thanks to the choice of \(j_m\) and Lemma A.1.
Thus we can identify \(\{v\}_C\) with the equivalence class \(\{(v_m)_{m\in \mathbb {N}}\}_{s,p}\). In other words, this proves the surjectivity of \(\mathcal {J}\). The proof is over. \(\square \)
Remark 4.5
We recall that we indicate
thus the functions belonging to this space are not necessarily bounded.
5 Characterisation for \(s\,p = n\)
5.1 General case
We start with the corresponding version Lemma 4.1 for the case \(s\,p=n\). The construction now is slightly more complicated, in particular it cannot be an easy consequence of scalings, as we have explained in the Introduction. Hence, a careful choice of auxiliary functions is needed.
Lemma 5.1
Let \(s \in (0,1]\) and \(n \ge 1\) be such that \(s<n\). There exists a sequence \((\varphi _m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) such that
as \(m\rightarrow +\infty \). Hence, \((\varphi _m)_{m\in \mathbb {N}}\) is equivalent in \(\mathcal {D}^{s,\frac{n}{s}} (\mathbb {R}^n)\) to zero, although its pointwise limit is 1.
Proof
As in [10, page 319] and [19, Lemma 15.2.2], we take the sequence
For \(s=1\), it can be checked by direct computation that
For the fractional case \(s\in (0,1)\), we claim that we still have
but the direct computation is fairly more intricate. We introduce the sequence
and observe that this is related to \(\psi _m\) through the relation
By using that the Gagliardo–Slobodeckiĭ seminorm is now scale invariant, we thus get
We now recall that from [25, Proposition 5.1] we have
We finally get that
as claimed.
Observe that technically speaking such a sequence \((\psi _m)_{m\in \mathbb {N}}\) does not belong to \(C^\infty _c (\mathbb {R}^n)\). However, this is a minor issue, that can be easily sorted by convolution. We take \(\rho \in C^\infty _0(\mathbb {R}^n)\) a standard Friedrichs mollifier supported on \(B_1(0)\), then by defining
we get the desired conclusion, thanks to the properties of convolutions. \(\square \)
The next technical result is the counterpart of Lemma 4.3 for the case \(s\,p=n\). In particular, the space of Hölder functions now has to be replaced by a suitable Campanato space.
Lemma 5.2
Let \(s \in (0,1]\) and \(n\ge 1\) be such that \(s<n\). Let \((u_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) be a Cauchy sequence with respect to \([\,\cdot \,]_{W^{s,n/s} (\mathbb {R}^n)}\). Then there exists another sequence \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\subset C_c^\infty (\mathbb {R}^n)\) such that:
-
we have
$$\begin{aligned} \lim _{m\rightarrow \infty }[u_m - {\widetilde{u}}_m]_{W^{s,\frac{n}{s}} (\mathbb {R}^n)}=0, \end{aligned}$$ -
\({\widetilde{u}}_m\) converges in \(L^\frac{n}{s}_\mathrm{loc}(\mathbb {R}^n)\) to a function \(u\in \mathcal {L}^{\frac{n}{s},n}(\mathbb {R}^n)\).
Moreover, this function u is such that
and we have
Proof
The proof is similar to that of Lemma 4.3. For every \(m\in \mathbb {N}\), we choose a natural number \(M_m\ge m\) large enough, so that
and consider the sequence \((\varphi _m)_{m\in \mathbb {N}}\) given by Lemma 5.1. By construction we have that \(M_m\) diverges to \(+\infty \), as m goes to \(\infty \). Then we define
where \(\varphi _{M_m}\) is as in Lemma 5.1.
It is not difficult to see that
We now show that \({\widetilde{u}}_m\) has the claimed properties. Thanks to the choice of \(M_m\) and Lemma 5.1, we have
which implies that
as desired. In particular, we get that \((\widetilde{u}_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) is still a Cauchy sequence with respect to the seminorm \([\,\cdot \,]_{W^{s,n/s} (\mathbb {R}^n)}\).
In order to infer the claimed convergence, we can apply Remark 2.6 with \(x_0=0\) and \(r=1\), so to get that \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\) is a Cauchy sequence in \(L^{n/s}(B_R)\), for every \(R\ge 1\). Thus we get that there exists \(u\in L^{n/s}_\mathrm{loc}(\mathbb {R}^n)\) such that
In particular, the strong convergence entails that
The first property is straightforward, by recalling (5.4). The fact that \(u\in \mathcal {L}^{\frac{n}{s},n}(\mathbb {R}^n)\) follows from
where we used Theorem 2.4. This permits to infer that
as claimed. Finally, the fact that
follows from the lower semicontinuity of the Gagliardo–Slobodeckiĭ seminorm with respect to the strong \(L^{n/s}\) convergence, which in turn follows from Fatou’s Lemma.
By using that \(({\widetilde{u}}_m)_{m\in \mathbb {N}} \subset C_c^\infty (\mathbb {R}^n)\) is still a Cauchy sequence with respect to \([\,\cdot \,]_{W^{s,p} (\mathbb {R}^n)}\) and arguing as in Part 2 of the proof of Theorem 3.1, we deduce (5.2). The estimate (5.3) can then be obtained by passing to the limit in the inequality of Theorem 2.4. \(\square \)
Theorem 5.3
Let \(s \in (0,1]\) and \(n \ge 1\) be such that \(s<n\). We define the quotient space
where \(\sim _C\) is the same equivalence relation as in (4.4). We still endow this space with the norm
Then \(\dot{W}^{s,\frac{n}{s} } (\mathbb {R}^n) \) is a Banach space and there exists a linear isometric isomorphism
In other words, the space \(\mathcal {D}^{s,\frac{n}{s}}(\mathbb {R}^n)\) can be identified with \(\dot{W}^{s,\frac{n}{s}}(\mathbb {R}^n)\).
Proof
The proof goes along the same lines of that of Theorem 4.4. The fact that \(\dot{W}^{s,n/s} (\mathbb {R}^n)\) is a normed vector space is straightforward.
Let us now consider the identification of our space. We take an equivalence class
and apply Lemma 5.2, so to get the new sequence \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\). By construction, we have
and by Lemma 5.2, we know that \(({\widetilde{u}}_m)_{m\in \mathbb {N}}\) converges to some function
Thus we may identify U with the equivalence class \(\{u\}_C\), i. e. we define \(\mathcal {J}(U) =\{u\}_C\). As in the case \(s\,p>n\), it is easily seen that this is a linear isometry.
On the other hand, for every equivalence class \(\{v\}_C\in \dot{W}^{s,p}(\mathbb {R}^n)\), we may select any representative v. Then we can construct a sequence \((v_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\) such that
It is indeed sufficient to repeat the construction of Part 1 in the proof of Theorem 3.1, with some minor modifications. We consider a sequence of cut-off functions \( \eta _j\in C^\infty _c ({\mathbb {R}}^n)\) with \(\mathrm {supp\,} \eta _j \subset B_{j^2}\) such that
and observe that by Lemma B.3, we have
Here we used the shortcut notation
Thus, for every \(m\ge 1\) we can choose \(j_m\in \mathbb {N}\) such that
The sequence \((v_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R}^n)\) is then given by
Indeed, by construction we have
Thus, by Lemma A.1 again, we get
as claimed. Accordingly, we can identify \(\{v\}_C\) with the equivalence class \(\{(v_m)_{m\in \mathbb {N}}\}_{s,p}\). In other words, this proves the surjectivity of \(\mathcal {J}\). The proof is complete.\(\square \)
Remark 5.4
In the case \(s\,p=n\), a natural replacement for the Sobolev inequality is the Moser-Trudinger inequality. However, this holds for open sets with finite measure, see e.g. [25].
5.2 The exceptional limit case \(s=n=1\)
Observe that in the previous section we have the restriction \(s<n\). Thus, in order to complete the picture in the conformal case, there is still a limiting case which is missing: the case \(s=1=n\). Accordingly, the summability exponent is \(p=1\), as well. This one-dimensional case is special and deserves to be treated separately.
As we will see, this situation is similar to the case \(s\,p < n\). Indeed, we have the following result.
Theorem 5.5
(The case \(s=p=n=1\)) Let us define
where the derivative \(u'\) is intended in the sense of distributions. Then there exists a linear isometric isomorphism
In other words, the space \(\mathcal {D}^{1,1}(\mathbb {R})\) can be identified with \(\dot{W}^{1,1}(\mathbb {R}^n)\).
Proof
By basic Calculus, we know that for every \(u\in C^\infty _c(\mathbb {R})\) and every \(z<x<y\) we have
and
By taking the limits as y goes to \(+\infty \) and z goes to \(-\infty \) in the previous inequalities, we get
By summing these two estimates and passing to supremum in x, we have an analogue to Sobolev’s inequality
Therefore, every Cauchy sequence \((u_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R})\) in the \(W^{1,1}\) seminorm is a Cauchy sequence in \(C_0(\mathbb {R})\), as well. The latter is the Banach space of continuous functions vanishing at infinity. Thus we can infer uniform convergence of \((u_m)_{m\in \mathbb {N}}\) to a function \(u\in C_0(\mathbb {R})\). Moreover, by using that \(L^1(\mathbb {R})\) is a Banach space, we can infer convergence of \((u'_m)_{m\in \mathbb {N}}\) to a function \(v\in L^1(\mathbb {R})\). It is easily seen that it must result
thus \(u\in \dot{W}^{1,1}(\mathbb {R})\). This argument permits to define the isometry \(\mathcal {J}\), exactly as we did in the proof of Theorem 3.1. In order to prove the surjectivity of \(\mathcal {J}\), it is sufficient to show that for every \(u\in \dot{W}^{1,1}(\mathbb {R})\), there exists a sequence \((u_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R})\) such that
This is a standard fact, we leave the details to the reader. \(\square \)
Remark 5.6
It is not difficult to see that inequality (5.5) is sharp. It is sufficient to take a sequence \((u_m)_{m\in \mathbb {N}}\subset C^\infty _c(\mathbb {R})\) such that
where u is the function
Such a sequence can be constructed by standard convolution methods.
6 Comments and open questions
1. Our three embedding theorems pose the question of what are the optimal constants, and whether they are achieved by extremal functions. We specifically refer to inequality (3.1) for \(s\,p<n\), inequality (2.9) for \(s\,p=n\), and inequality (2.11) for \(s\,p>n\).
Note that in the case \(s=1\), the extremals in the Gagliardo–Nirenberg–Sobolev range \(1<p<n\) were found by Aubin [2] and Talenti [28]. The extremals in the Morrey range \(p>n\), still for \(s=1\), have been recently described by Hynd and Sauffert [15]. We do not know of any similar result in the limit case \(p=n\).
2. In the case of proper subsets \(\Omega \subset \mathbb {R}^n\) the characterization result will depend on the different options. Thus, one may take completions of \(C^\infty _c(\Omega )\) with respect to:
-
the full norm (but localized on \(\Omega \))
$$\begin{aligned} \Vert u\Vert _{L^p(\Omega )}+[u]_{W^{s,p}(\Omega )}; \end{aligned}$$ -
the full norm (but spread all over \(\mathbb {R}^n\)),
$$\begin{aligned} \Vert u\Vert _{L^p(\Omega )}+[u]_{W^{s,p}(\mathbb {R}^n)}; \end{aligned}$$ -
the Gagliardo–Slobodeckiĭ seminorm (localized on \(\Omega \))
$$\begin{aligned}{}[u]_{W^{s,p}(\Omega )}; \end{aligned}$$ -
or the Gagliardo–Slobodeckiĭ seminorm (spread all over \(\mathbb {R}^n\)).
In general, the resulting spaces do not coincide. See [6, Section 2] for some comments.
Notes
We can use the following fact: if we set
$$\begin{aligned} f(\tau )=\frac{\tau }{(1+\tau ^\frac{2}{\alpha })^\frac{\alpha }{2}},\quad \hbox { for } \tau \ge 0, \end{aligned}$$then this is a Lipschitz function and we have
$$\begin{aligned} \varphi (x)=(1+|x|^2)^{-\frac{\alpha }{2}}=f(|x|^{-\alpha }). \end{aligned}$$
References
Adams, R., Fournier, J.: Sobolev Spaces (Pure and Applied Mathematics), vol. 140, 2nd edn. Elsevier, Amsterdam (2003)
Aubin, T.: Problèmes isopérimétriques et espaces de Sobolev. J. Differ. Geom. 11, 573–598 (1976)
Bergh, J., Löfström, J.: Interpolation Spaces. An introduction: Grundlehren der Mathematischen Wissenschaften, vol. 223. Springer, Berlin (1976)
Bourgain, J., Brezis, H., Mironescu, P.: Another look at Sobolev spaces. In: Menaldi, J.L., Rofman, E., Sulem, A. (eds.) Optimal Control and Partial Differential Equations. A Volume in Honor of Professor Alain Bensoussan’s 60th Birthday, pp. 439–455. IOS Press, Amsterdam (2001)
Brasco, L., Mosconi, S., Squassina, M.: Optimal decay of extremals for the fractional Sobolev inequality. Calc. Var. Partial Differ. Equ. 55, Paper No. 23, 32 pp (2016)
Brasco, L., Salort, A.: A note on homogeneous Sobolev spaces of fractional order. Ann. Mat. Pura Appl. (4) 198, 1295–1330 (2019)
Campanato, S.: Proprietà di hölderianità di alcune classi di funzioni. Ann. Scuola Norm. Sup. Pisa (3) 17, 175–188 (1963)
Cotsiolis, A., Tavoularis, N.: Best constants for Sobolev inequalities for higher order fractional derivatives. J. Math. Anal. Appl. 295, 225–236 (2004)
del Teso, F., Gómez-Castro, D., Vázquez, J.L.: Estimates on translations and Taylor expansions in fractional Sobolev spaces. Nonlinear Anal. 200, 111995 (2020)
Deny, J., Lions, J.-L.: Les espaces du type de Beppo-Levi. Ann. Inst. Fourier 5, 305–370 (1954)
Fefferman, C.: Characterizations of bounded mean oscillation. Bull. Am. Math. Soc. 77, 587–588 (1971)
Fefferman, C., Stein, E.M.: \(H^p\) spaces of several variables. Acta Math. 129, 137–193 (1972)
Franzina, G.: Non-local torsion functions and embeddings. Appl. Anal. 98, 1811–1826 (2019)
Giusti, E.: Direct Methods in the Calculus of Variations. World Scientific Publishing Co. Inc., River Edge (2003)
Hynd, R., Seuffert, F.: Extremal Functions for Morrey’s Inequality (2018) (Preprint). arxiv:1810.04393
John, F., Nirenberg, L.: On functions of bounded mean oscillation. Commun. Pure Appl. Math. 14, 415–426 (1961)
Leoni, G.: A First Course in Sobolev Spaces (Graduate Studies in Mathematics), vol. 181, 2nd edn. American Mathematical Society, Providence (2017)
Lions, J.-L.: Théorèmes de trace et d’interpolation. I. Ann. Scuola Norm. Sup. Pisa (3) 13, 389–403 (1959)
Maz’ya, V.: Sobolev Spaces with Applications to Elliptic Partial Differential Equations. Second revised and augmented edition. Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], vol. 342. Springer, Heidelberg (2011)
Maz’ya, V., Shaposhnikova, T.: On the Bourgain, Brezis, and Mironescu theorem concerning limiting embeddings of fractional Sobolev spaces. J. Funct. Anal. 195, 230–238 (2002)
Mingione, G.: The Singular Set of Solutions to Non-Differentiable Elliptic Systems. Arch. Rational Mech. Anal., 166, 287–301 (2003)
Mironescu, P., Russ, E.: Traces of weighted Sobolev spaces. Old and new. Nonlinear Anal. 119, 354–381 (2015)
Monguzzi, A., Peloso, M.M., Salvatori, M.: Fractional Laplacian, homogeneous Sobolev spaces and their realizations. Ann. Mat. Pura Appl. (4) 199, 2243–2261 (2020)
Moser, J.: On Harnack’s theorem for elliptic differential equations. Commun. Pure Appl. Math. 16, 577–591 (1961)
Parini, E., Ruf, B.: On the Moser–Trudinger inequality in fractional Sobolev–Slobodeckij spaces. J. Anal. Math. 138, 281–300 (2019)
Peetre, J.: Espaces d’interpolation et théorème de Soboleff. Ann. Inst. Fourier (Grenoble) 16, 279–317 (1966)
Stein, E.M.: Harmonic Analysis: Real-Variable Methods, Orthogonality, and Oscillatory Integrals: Princeton Mathematical Series. Monographs in Harmonic Analysis, III, vol. 43. Princeton University Press, Princeton (1993)
Talenti, G.: Best constant in Sobolev inequality. Ann. Mat. Pura Appl. (4) 119, 353–372 (1976)
Tartar, L.: An Introduction to Sobolev Spaces and Interpolation Spaces. Lecture Notes of the Unione Matematica Italiana, vol. 3. Springer, Berlin (2007)
Triebel, H.: Theory of Function Spaces. Monographs in Mathematics, vol. 78. Birkhäuser, Basel (1983)
Vázquez, J.L.: The Dirichlet Problem for the fractional \(p\)-Laplacian evolution equation. J. Differ. Equ. 260(7), 6038–6056 (2016)
Vázquez, J.L.: The evolution fractional \(p\)-Laplacian equation in \(\mathbb{R}^n\). Fundamental solution and asymptotic behaviour. Nonlinear Anal. 199, 112034 (2000)
Warma, M.: The fractional relative capacity and the fractional Laplacian with Neumann and Robin boundary conditions on open sets. Potential Anal. 42, 499–547 (2015)
Acknowledgements
An anonymous referee is gratefully acknowledged for the careful reading of the manuscript and for the many useful comments. We also wish to thank Alessandro Monguzzi for having kindly drawn our attention to the paper [23]. L. Brasco was financially supported by the grant FFABR Fondo Per il Finanziamento delle attività di base, of the Italian Government. The work of D. Gómez-Castro and J. L. Vázquez were funded by grant PGC2018-098440-B-I00 from the MICINN, of the Spanish Government. The research of D. Gómez-Castro was supported by the Advanced Grant Nonlocal – CPD (Nonlocal PDEs for Complex Particle Dynamics: Phase Transitions, Patterns and Synchronization) of the European Research Council Executive Agency (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No. 883363). J. L. Vázquez is an Honorary Professor at Universidad Complutense de Madrid.
Funding
Open access funding provided by Università degli Studi di Ferrara within the CRUI-CARE Agreement.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by A. Malchiodi.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A. Approximation by convolution
Let \(\rho \in C^\infty _c(\mathbb {R}^n)\) be a standard Friedrichs mollifier supported on the ball \(B_1(0)\), and define the sequence of smoothing kernels
In [9, Theorem 2.3], the second and third authors proved, by means of interpolation techniques, that
for a constant \(C=C(n)>0\). Here, we will need a stronger result (indeed, our functions need not belong to \(W^{s,p}(\mathbb {R}^n)\)), but without a rate of convergence. The proof is standard routine in the theory of \(L^p\) spaces, we recall the argument for the reader’s convenience.
Lemma A.1
Let \(s\in [0,1]\) and \(1\le p<+\infty \). For every \(u\in L^1_\mathrm{loc}(\mathbb {R}^n)\) such that
we have
Proof
We focus on the case \(s\in (0,1)\), the extremal cases \(s=0\) and \(s=1\) being simpler and well-known. By using that
we have
By using Jensen’s inequality and the fact that \(\rho _m(z)\,\rho _m(w)\,dz\,dw\) is a probability measure, we then get
If we now exchange the order of integration, use that \(\rho _m\) is supported on \(B_{1/m}(0)\) and the fact that
we easily get the desired conclusion (A.1). \(\square \)
Appendix B. Truncation lemmas
In what follows, we will use the shortcut notation \(B_r\) for \(B_r(0)\). The aim of this section is to show that there exist smooth cut-off functions \(\eta _m\) with
for suitable radii \(a_m \le b_m\) diverging to \(+\infty \), such that for every \([u]_{W^{s,p} ({\mathbb {R}}^n) } < +\infty \) we have
We will prove this for \(s\,p < n\) and \(s\,p = n\). For the case \(s\,p > n\), we will only show that \([ \eta _m \, u - u ]_{W^{s,p} (\mathbb R^n) }\) is uniformly bounded: this is sufficient for our scope, since we can then apply weak compactness and a convexity trick based on Mazur’s Lemma.
For convenience, we will write \(\varphi _m = 1 - \eta _m\), and prove that
The main difficulty arises from the fact that only minimal integrability assumptions are assumed on u. In particular, we do not require that \(u\in W^{s,p}(\mathbb {R}^n)\).
1.1 B.1. Case \(s\,p < n\)
For \(0<s\le 1\) and \(1\le p<+\infty \) such that \(s\,p<n\), we recall the definition of critical Sobolev exponent
Then we have the following technical result, which is quite useful.
Lemma B.1
(Truncation lemma \(s\,p<n\)) Let
If \((\varphi _m)_{m\in \mathbb {N}}\subset C^\infty _0(\mathbb {R}^n)\) is a sequence of non-negative cut-off functions such that \(0\le \varphi _m\le 1\) and
then we have
Proof
We deal with the case \(s\in (0,1)\), the case \(s=1\) being much simpler. We decompose the seminorm as follows
We show that each \(\mathcal {I}_i\) converges to 0. We treat each integral separately.
Estimate for \(\mathcal {I}_1\). For the first integral, we first observe that
By noticing that
we get
and thus
The last integral can be estimated by Hölder’s inequality
By observing that
we conclude that \(\mathcal {I}_1\) converges to 0 as m goes to \(\infty \), by using that \(u\in L^{p^\star _s}(\mathbb {R}^n)\) and the Dominated Convergence Theorem.
Estimate for \(\mathcal {I}_2\). As for \(\mathcal {I}_2\), by using Fubini’s Theorem and Hölder’s inequality, we get
Observe that
thus we get
We then obtain
Since we have \(u\in L^{p^\star _s}(\mathbb {R}^n)\), the last integral converges to 0, as m goes to 0.
Estimate for \(\mathcal {I}_3\). For the third integral, we have
Now, the second integral converges to 0 by the Dominated Convergence Theorem. The first one can be handled as we did for \(\mathcal {I}_1\).
Estimate for \(\mathcal {I}_4\). We observe that
where we used that \(\varphi _m=1\) on the complement of \(B_{2\,m}\). The last integral converges to 0, while for the first one we further decompose it as follows
For the first integral, we can observe that, if \(x \in B_{2\,m}{\setminus } B_m\), then
hence \(\mathbb {R}^n{\setminus } B_{3\,m} \subset {\mathbb {R}}^n {\setminus } B_{m}(x)\). This yields
The last term converges to 0, as we have shown while estimating \(\mathcal {I}_1\). For the other remaining integral, we can use a similar estimate as for \(\mathcal {I}_1\). We leave the details to the reader.
Estimate for \(\mathcal {I}_5\). This is the simplest term, we just observe that
by the Dominated Convergence Theorem. This concludes the proof. \(\square \)
1.2 B.2. Case \(s\,p > n\)
In this case, the estimate on the truncation will be slightly worse. However, this is still sufficient for our scopes.
Lemma B.2
(Truncation lemma \(s\,p>n\)) Let \(s\in (0,1]\) and \(1\le p<+\infty \) be such that \(s\,p>n\). We set \(\alpha =s-n/p\) and let
If \((\varphi _m)_{m\in \mathbb {N}}\subset C^\infty _0(\mathbb {R}^n)\) is a sequence of non-negative cut-off functions such that \(0\le \varphi _m\le 1\) and
then we have
for a constant \(C>0\) not depending on m.
Proof
The proof is similar to the previous one. We deal again with the case \(s\in (0,1)\). We still decompose the \((s,p)-\)seminorm as before and then use the estimate
in place of the hypothesis \(L^{p^\star _s}\) previously used, in order to estimate \(\mathcal {I}_1\), \(\mathcal {I}_3\) and \(\mathcal {I}_4\). We leave the details to the reader. For \(\mathcal {I}_2\), we observe that
We then use that
Thus we obtain
This concludes the proof. \(\square \)
1.3 B.3. Case \(s\,p = n\)
Here we can prove a result similar to Lemma B.1. For this, we will need the Poincaré–Wirtinger inequality of Lemma 2.9 and the integrability information of Lemma 2.2.
Lemma B.3
(Truncation lemma \(s\,p=n)\) Let \(s\in (0,1]\) and \(n\ge 1\) be such that \(s<n\). Let
If \((\varphi _m)_{m\in \mathbb {N}}\subset C^\infty _0(\mathbb {R}^n)\) is a sequence of non-negative cut-off functions such that \(0\le \varphi _m\le 1\)
then we have
Proof
We only consider the case \(s\in (0,1)\), the local case being much simpler. We decompose the seminorm as usual
We treat each integral separately.
Estimate for \(\mathcal {I}_1\). For the term \(\mathcal {I}_1\), we first observe that for every \(y\in B_{m^2}{\setminus } B_m\) and every \(x\in B_m\), we have
This entails
By noticing that
we get
and thus
If we now apply Lemma 2.9 on the right-hand side, we get
By using that \([u]_{W^{s,n/s}(\mathbb {R}^n)}<+\infty \) and the Dominated Convergence Theorem, we get that the last term above converges to 0, as m goes to \(\infty \).
Estimate for \(\mathcal {I}_2\). This term now is quite delicate, here we need a global integrability information on u. We observe that for every \(m\ge 2\), \(x\in B_m\) and \(y\in \mathbb {R}^n{\setminus } B_{m^2}\), we have
Thus we get
We now observe that for \(m\ge 2\), \(|y|\ge m^2\) and \(0<\alpha <n/2\), we have
and also
for a suitable constant \(C_\alpha >1\). By combining the two previous estimates, we then obtain
We can then estimate the integral in the right-hand side of (B.1) as follows
By using Lemma 2.2 with \(p=n/s\), we get that \(\mathcal {I}_2\) converges to 0, as m goes to \(\infty \).
Estimate for \(\mathcal {I}_3\). For the third integral, we have
Now, the second integral converges to 0 by the Dominated Convergence Theorem. The first one can be handled as we did for \(\mathcal {I}_1\).
Estimate for \(\mathcal {I}_4\). Here, we observe that
where we used that \(\varphi _m=1\) on the complement of \(B_{m^2}\). The last integral converges to 0, while the first one can be decomposed as follows:
For the first integral, we can observe that, if \(x \in B_{m^2}{\setminus } B_m\), then
hence \(\mathbb {R}^n{\setminus } B_{2\,m^2} \subset {\mathbb {R}}^n {\setminus } B_{m^2}(x)\). This yields
Proceeding as for \(\mathcal {I}_1\), the last term converges to 0.
For the second integral we use a similar estimate as for \(\mathcal {I}_1\). More precisely, we observe that for \(x\in B_{m^2}\), we have \( B_{2\,m^2}{\setminus } B_{m^2}\subset B_{3\,m^2}(x). \) Thus we can estimate
The latter is again the same term previously treated.
Estimate for \(\mathcal {I}_5\). As usual, this is the simplest term, it results
by the Dominated Convergence Theorem. The desired conclusion now follows. \(\square \)
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Brasco, L., Gómez-Castro, D. & Vázquez, J.L. Characterisation of homogeneous fractional Sobolev spaces. Calc. Var. 60, 60 (2021). https://doi.org/10.1007/s00526-021-01934-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00526-021-01934-6