Factorization in Lorentz spaces, with an application to centralizers (Q333875)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Factorization in Lorentz spaces, with an application to centralizers |
scientific article |
Statements
Factorization in Lorentz spaces, with an application to centralizers (English)
0 references
31 October 2016
0 references
A classical result by Lozanovsky provides a factorization for any function belonging to \(L^1\) as a product of two functions, one of them belonging to a function space \(X\) and the other one to its Köthe dual in a way that the product of the norms is (almost) equal to the norm of \(f\) in \(L^1\). In the same direction of this fundamental result, we can find in the mathematical literature several relevant theorems providing factorizations for functions belonging to other function spaces and giving some control on their norms. However, very often -- beyond the case of the \(L^p\)-spaces -- it is not easy to explicitly compute such factorization. The present paper provides this factorization for the case of Lorentz spaces. The main theorem gives the following Theorem. Consider extended real numbers \(0<p_0,p_1,q_0,q_1,p,q\) and the Lorentz spaces \(L(p_0,q_0),\) \(L(p_1,q_1)\) and \(L(p,q)\). Suppose that \( (p_0,q_0)^{-1} + (p_1,q_1)^{-1}=(p,q)^{-1}\). Then there is a constant \(M\) depending only on \(p\) and \(q\) such that {\parindent=0.7cm \begin{itemize}\item[(a)] If \(f_i \in L(p_i,q_i)\), \(i=0,1\), then \(f_0 f_1 \in L(p,q)\) and \[ \|f_0 f_1 \|_{p,q} \leq M \|f_0\|_{p_0,q_0} \, \|f_1\|_{p_1,q_1}. \] \item[(b)] If \(f \in L(p,q)\), then there are \(f_i \in L((p_i,q_i)\), \(i=0,1\), such that \(f=f_0 f_1\) and \[ \|f_0\|_{p_0,q_0} \, \|f_1\|_{p_1,q_1} \leq M \|f \|_{p,q}. \] \item[(c)] If \( q< \infty\) and \(f \geq 0\), one can take \(f_i= f^{q q_i^{-1}} \, r_f^{qp^{-1} q_i^{-1}- p_i^{-1}}\) for \(i=0,1\) in (b), where \(r_f\) is the rank function for \(f\). For \(q= \infty\) and \(f \geq 0\), one may take \(f_i=f^{p p_i^{-1}}\) in (b). \end{itemize}} Given a function \(f\), the rank function \(r_f\) is given by \[ r_f(t) = m \{s: |f(s)| > |f(t)| \, \text{or}\, |f(s)|= |f(t)| \, \text{and}\, s \leq t \}. \] Some applications to centralizers are also given.
0 references
Lorentz space
0 references
Lozanovsky factorization
0 references
centralizer
0 references
twisted sum
0 references
interpolation theory
0 references