We present the complex interpolation of Besov and Triebel–Lizorkin spaces with generalized smoothness. In some particular cases these function spaces are just weighted Besov and Triebel–Lizorkin spaces. As a corollary of our results, we obtain the complex interpolation between the weighted Triebel–Lizorkin spaces ${\dot{F}}_{p_0,q_0}^{s_0}\left({\omega }_0\right)$ and ${\dot{F}}_{\infty ,q_1}^{s_1}\left({\omega }_1\right)$ with suitable assumptions on the parameters $s_0,s_1,p_0,q_0$ and $q_1$, and the pair of weights $({\omega }_0,{\omega }_1)$.

If C is a capacity on a measurable space, we prove that the restriction of the K-functional $K(t,f;L^p\left(C\right),L^{\infty }\left(C\right))$ to quasicontinuous functions f ∈ QC is equivalent to $K(t,f;L^p\left(C\right)\cap QC,L^{\infty }\left(C\right)\cap QC)$. We apply this result to identify the interpolation space ${(L^{p₀,q₀}\left(C\right)\cap QC,L^{p₁,q₁}\left(C\right)\cap QC)}_{\theta ,q}$.

The interpolation properties of Cesàro sequence and function spaces are investigated. It is shown that $Ce{s}_p\left(I\right)$ is an interpolation space between $Ce{s}_{p₀}\left(I\right)$ and $Ce{s}_{p₁}\left(I\right)$ for 1 < p₀ < p₁ ≤ ∞ and 1/p = (1 - θ)/p₀ + θ/p₁ with 0 < θ < 1, where I = [0,∞) or [0,1]. The same result is true for Cesàro sequence spaces. On the other hand, $Ce{s}_p[0,1]$ is not an interpolation space between Ces₁[0,1] and $Ce{s}_{\infty }[0,1]$.

A complete description of the real interpolation space $L={(L_{p₀}\left(\omega ₀\right),...,L_{pₙ}\left(\omega ₙ\right))}_{\theta ⃗,q}$ is given. An interesting feature of the result is that the whole measure space (Ω,μ) can be divided into disjoint pieces ${\Omega }_i$ (i ∈ I) such that L is an $l_q$ sum of the restrictions of L to ${\Omega }_i$, and L on each ${\Omega }_i$ is a result of interpolation of just two weighted $L_p$ spaces. The proof is based on a generalization of some recent results of the first two authors concerning real interpolation of vector-valued spaces.

We study imbeddings of the Sobolev space $W^{m,\varrho }\left(\Omega \right)$: = u: Ω → ℝ with $\varrho ({\partial }^{\alpha }u/\partial x^{\alpha })$ < ∞ when |α| ≤ m, in which Ω is a bounded Lipschitz domain in ℝⁿ, ϱ is a rearrangement-invariant (r.i.) norm and 1 ≤ m ≤ n - 1. For such a space we have shown there exist r.i. norms, ${\tau }_{\varrho }$ and ${\sigma }_{\varrho }$, that are optimal with respect to the inclusions $W^{m,\varrho }\left(\Omega \right)\subset W^{m,{\tau }_{\varrho }}\left(\Omega \right)\subset L_{{\sigma }_{\varrho }}\left(\Omega \right)$. General formulas for ${\tau }_{\varrho }$ and ${\sigma }_{\varrho }$ are obtained using the -method of interpolation. These lead to explicit expressions when ϱ is a Lorentz Gamma norm or an Orlicz norm.

We find necessary and sufficient conditions on a pair of rearrangement-invariant norms, ϱ and σ, in order that the Sobolev space $W^{m,\varrho }\left(\Omega \right)$ be compactly imbedded into the rearrangement-invariant space $L_{\sigma }\left(\Omega \right)$, where Ω is a bounded domain in ℝⁿ with Lipschitz boundary and 1 ≤ m ≤ n-1. In particular, we establish the equivalence of the compactness of the Sobolev imbedding with the compactness of a certain Hardy operator from $L_{\varrho }(0,|\Omega \left|\right)$ into $L_{\sigma }(0,|\Omega \left|\right)$. The results are illustrated with examples in which ϱ and σ are both...

This paper deals with Besov spaces of logarithmic smoothness $B_{p,r}^{0,b}$ formed by periodic functions. We study embeddings of $B_{p,r}^{0,b}$ into Lorentz-Zygmund spaces $L_{p,q}{\left(logL\right)}_{\beta }$. Our techniques rely on the approximation structure of $B_{p,r}^{0,b}$, Nikol’skiĭ type inequalities, extrapolation properties of $L_{p,q}{\left(logL\right)}_{\beta }$ and interpolation.

We find necessary and sufficient conditions under which the norms of the interpolation spaces ${(N₀,N₁)}_{\theta ,q}$ and ${(X₀,X₁)}_{\theta ,q}$ are equivalent on N, where N is the kernel of a nonzero functional ψ ∈ (X₀ ∩ X₁)* and $N_i$ is the normed space N with the norm inherited from $X_i$ (i = 0,1). Our proof is based on reducing the problem to its partial case studied by Ivanov and Kalton, where ψ is bounded on one of the endpoint spaces. As an application we completely resolve the problem of when the range of the operator $T_{\theta }=S-2^{\theta }I$ (S...

We prove the existence of solutions to the evolutionary Stokes system in a bounded domain Ω ⊂ ℝ³. The main result shows that the velocity belongs either to $W_p^{2s+2,s+1}\left({\Omega }^T\right)$ or to $B_{p,q}^{2s+2,s+1}\left({\Omega }^T\right)$ with p > 3 and s ∈ ℝ₊ ∪ 0. The proof is divided into two steps. First the existence in $W_p^{2k+2,k+1}$ for k ∈ ℕ is proved. Next applying interpolation theory the existence in Besov spaces in a half space is shown. Finally the technique of regularizers implies the existence in a bounded domain. The result is generalized to the spaces...

The purpose of this paper is to provide a new characterization of the Sobolev space $W^{1,1}\left(ℝⁿ\right)$. We also show a new proof of the characterization of the Sobolev space $W^{1,p}\left(ℝⁿ\right)$, 1 ≤ p < ∞, in terms of Poincaré inequalities.

Assume ||·|| is a norm on ℝⁿ and ||·||⁎ its dual. Consider the closed ball $T:=B_{\left|\right|·\left|\right|}(0,r)$, r > 0. Suppose φ is an Orlicz function and ψ its conjugate. We prove that for arbitrary A,B > 0 and for each Lipschitz function f on T, $su{p}_{s,t\in T}|f\left(s\right)-f\left(t\right)|\le 6AB({\int }_0^r\psi (1/A{\epsilon }^{n-1}){\epsilon }^{n-1}d\epsilon +1/\left(n\right|B_{\left|\right|·\left|\right|}(0,1)\left|\right){\int }_T\phi (1/B||\nabla f\left(u\right)\left|\right|⁎\left)du\right)$, where |·| is the Lebesgue measure on ℝⁿ. This is a strengthening of the Sobolev inequality obtained by M. Talagrand. We use this inequality to state, for a given concave, strictly increasing function η: ℝ₊ → ℝ with η(0) = 0, a necessary and sufficient condition on...

Let Ω,Ω’ ⊂ ℝⁿ be domains and let f: Ω → Ω’ be a homeomorphism. We show that if the composition operator $T_f:u\mapsto u\circ f$ maps the Sobolev-Lorentz space $W{L}^{n,q}\left({\Omega }^{\text{'}}\right)$ to $W{L}^{n,q}\left(\Omega \right)$ for some q ≠ n then f must be a locally bilipschitz mapping.

For a function $f\in L_{loc}^p\left(ℝⁿ\right)$ the notion of p-mean variation of order 1, ${₁}^p(f,ℝⁿ)$ is defined. It generalizes the concept of F. Riesz variation of functions on the real line ℝ¹ to ℝⁿ, n > 1. The characterisation of the Sobolev space $W^{1,p}\left(ℝⁿ\right)$ in terms of ${₁}^p(f,ℝⁿ)$ is directly related to the characterisation of $W^{1,p}\left(ℝⁿ\right)$ by Lipschitz type pointwise inequalities of Bojarski, Hajłasz and Strzelecki and to the Bourgain-Brezis-Mironescu approach.

Let $\theta \in (0,1)$, $\lambda \in [0,1)$ and $p,p_0,p_1\in (1,\infty ]$ be such that $(1-\theta )/p_0+\theta /p_1=1/p$, and let $\varphi ,{\varphi }_0,{\varphi }_1$ be some admissible functions such that $\varphi ,{\varphi }_0^{p/p_0}$ and ${\varphi }_1^{p/p_1}$ are equivalent. We first prove that, via the $\pm$ interpolation method, the interpolation $\langle L_{{\varphi }_0}^{p_0),\lambda }\left(𝒳\right),L_{{\varphi }_1}^{p_1),\lambda }\left(𝒳\right),\theta \rangle$ of two generalized grand Morrey spaces on a quasi-metric measure space $𝒳$ is the generalized grand Morrey space $L_{\varphi }^{p),\lambda }\left(𝒳\right)$. Then, by using block functions, we also find a predual space of the generalized grand Morrey space. These results are new even for generalized grand Lebesgue spaces.

The well-known Bohr-Pál theorem asserts that for every continuous real-valued function f on the circle there exists a change of variable, i.e., a homeomorphism h of onto itself, such that the Fourier series of the superposition f ∘ h converges uniformly. Subsequent improvements of this result imply that actually there exists a homeomorphism that brings f into the Sobolev space $W{₂}^{1/2}\left(\right)$. This refined version of the Bohr-Pál theorem does not extend to complex-valued functions. We show that if...

We establish an interpolation theorem for a class of nonlinear operators in the Lebesgue spaces ${ℒ}^s\left(ℝⁿ\right)$ arising naturally in the study of elliptic PDEs. The prototype of those PDEs is the second order p-harmonic equation ${div\left|\nabla u\right|}^{p-2\nabla }u=div$. In this example the p-harmonic transform is essentially inverse to $div\left(\right|\nabla {|}^{p-2}\nabla )$. To every vector field $\in {ℒ}^q(ℝⁿ,ℝⁿ)$ our operator ${\mathscr{H}}_p$ assigns the gradient of the solution, ${\mathscr{H}}_p=\nabla u\in {ℒ}^p(ℝⁿ,ℝⁿ)$. The core of the matter is that we go beyond the natural domain of definition of this operator. Because of nonlinearity our...

We study embeddings of spaces of Besov-Morrey type, $i{d}_{\Omega }{:}_{p₁,u₁,q₁}^{s₁}\left(\Omega \right){\hookrightarrow }_{p₂,u₂,q₂}^{s₂}\left(\Omega \right)$, where $\Omega \subset {ℝ}^d$ is a bounded domain, and obtain necessary and sufficient conditions for the continuity and compactness of $i{d}_{\Omega }$. This continues our earlier studies relating to the case of ${ℝ}^d$. Moreover, we also characterise embeddings into the scale of $L_p$ spaces or into the space of bounded continuous functions.

The estimate ${∥D^{k-1}u∥}_{L^{n/(n-1)}}\le {∥A\left(D\right)u∥}_{L^1}$ is shown to hold if and only if $A\left(D\right)$ is elliptic and canceling. Here $A\left(D\right)$ is a homogeneous linear differential operator $A\left(D\right)$ of order $k$ on ${ℝ}^n$ from a vector space $V$ to a vector space $E$. The operator $A\left(D\right)$ is defined to be canceling if ${\bigcap }_{\xi \in {ℝ}^n\setminus \left\{0\right\}}A\left(\xi \right)\left[V\right]=\left\{0\right\}$. This result implies in particular the classical Gagliardo–Nirenberg–Sobolev inequality, the Korn–Sobolev inequality and Hodge–Sobolev estimates for differential forms due to J. Bourgain and H. Brezis. In the proof, the class of cocanceling homogeneous...

Let $(A_0,A_1)$ be a regular interpolation couple. Under several different assumptions on a fixed $A^{\beta }$, we show that $A^{\theta }=A_{\theta }$ for every $\theta \in (0,1)$. We also deal with assumptions on ${\overline{A}}^{\beta }$, the closure of $A^{\beta }$ in the dual of ${(A_0^{*},A_1^{*})}_{\beta }$.

Given a compact manifold $N^n$, an integer $k\in {\mathbb{N}}_{*}$ and an exponent $1\le p<\infty$, we prove that the class $C^{\infty }({\overline{Q}}^m;N^n)$ of smooth maps on the cube with values into $N^n$ is dense with respect to the strong topology in the Sobolev space $W^{k,p}(Q^m;N^n)$ when the homotopy group ${\pi }_{\lfloor kp\rfloor }\left(N^n\right)$ of order $\lfloor kp\rfloor$ is trivial. We also prove density of maps that are smooth except for a set of dimension $m-\lfloor kp\rfloor -1$, without any restriction on the homotopy group of $N^n$.

Let A be a linear closed densely defined operator in a complex Banach space X. If A is of type ω (i.e. the spectrum of A is contained in a sector of angle 2ω, symmetric around the real positive axis, and $\parallel \lambda {(\lambda I-A)}^{-1}\parallel$ is bounded outside every larger sector) and has a bounded inverse, then A has a bounded $H^{\infty }$ functional calculus in the real interpolation spaces between X and the domain of the operator itself.