Grothendieck-type characterization of representable multilinear mappings

DOI10.1080/03081087.2024.2352776MaRDI QIDQ7030484zbMATH OpenFDO

Authors Raffaella Cilia, Joaquín M. Gutiérrez

Publication date 7 February 2025

Published in Linear and Multilinear Algebra (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1080/03081087.2024.2352776

(Spaces of) multilinear mappings, polynomials (46G25) Linear operators belonging to operator ideals (nuclear, (p)-summing, in the Schatten-von Neumann classes, etc.) (47B10) Multilinear and polynomial operators (47H60)

Review text

Let \(k\) be a natural number, \(\left( \Omega _{i},\Sigma _{i},\mu _{i}\right) _{1\leq i\leq k}\) finite measure spaces and \(\left( \Omega ,\Sigma ,\mu \right) \) be the product of these measure spaces. A \(k\)-linear mapping \( A:L_{1}\left( \mu _{1}\right) \times \cdot \cdot \cdot \times L_{1}\left( \mu _{k}\right) \rightarrow X\) is representable if there is \(g\in L_{\infty }\left( \mu ,X\right) \) such that \(A\left( f_{1},\dots ,f_{k}\right) =\int_{\Omega }f_{1}\left( t_{1}\right) \cdot \cdot \cdot f_{k}\left( t_{k}\right) g\left( t_{1},\dots ,t_{k}\right) d\mu \) for all \(f_{1}\in L_{1}\left( \mu _{1}\right) \), \dots , \(f_{k}\in L_{1}\left( \mu _{k}\right) \). The main result of the paper is Theorem~4.1:\N\N For every \(1\leq i\leq k\) let \(\left( \Omega _{i},\Sigma _{i},\mu _{i}\right) _{1\leq i\leq k}\) be a finite measure space and \(I_{i}:L_{\infty }\left( \mu _{i}\right) \rightarrow L_{1}\left( \mu _{i}\right) \) be the natural inclusion. For a \(k\)-linear mapping \(A:L_{1}\left( \mu _{1}\right) \times \cdot \cdot \cdot \times L_{1}\left( \mu _{k}\right) \rightarrow X\) the following assertions are equivalent:\N\begin{itemize}\N\item[(a)] \N\(A\) is representable;\N\item[(b)] \Nthe \(k\)-linear mapping\N\[\NL_{\infty }\left( \mu _{1}\right) \times \cdot \cdot \cdot \times L_{\infty }\left( \mu _{k}\right) \overset{I_{1}\times \cdot \cdot \cdot \times I_{k}}{ \rightarrow }L_{1}\left( \mu _{1}\right) \times \cdot \cdot \cdot \times L_{1}\left( \mu _{k}\right) \overset{A}{\rightarrow }X\N\]\Nis nuclear.\N\end{itemize}\NMoreover, \N\[\N\left\Vert A\circ \left( I_{1}\times \cdot \cdot \cdot \times I_{k}\right) \right\Vert _{\mathrm{nuc}}\N=\int_{\Omega }\left\Vert g\left( \omega _{1},...,\omega _{k}\right) \right\Vert d\mu _{1}\cdot \cdot \cdot d\mu _{k}=\left\Vert \overline{A\circ \left( I_{1}\times \cdot \cdot \cdot \times I_{k}\right) } \right\Vert _{\mathrm{nuc}},\N\]\Nwhere \(g\) is the density of \(A\) and \(\overline{ A\circ \left( I_{1}\times \cdot \cdot \cdot \times I_{k}\right) }:L_{\infty }\left( \mu _{1}\right) \widehat{\otimes }_{\varepsilon }\cdot \cdot \cdot \widehat{\otimes }_{\varepsilon }L_{\infty }\left( \mu _{k}\right) \rightarrow X\) is its linearization.\N\NAn application in the case of polynomials is given.

Cites work

Cited in

(2)

This page was built for publication: Grothendieck-type characterization of representable multilinear mappings

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7030484)