Isomorphic embedding of \(\ell_p^n\), \(1<p<2\), into \(\ell_1^{(1+\varepsilon)n}\) (Q5935818)

From MaRDI portal
scientific article; zbMATH DE number 1611101
Language Label Description Also known as
English
Isomorphic embedding of \(\ell_p^n\), \(1<p<2\), into \(\ell_1^{(1+\varepsilon)n}\)
scientific article; zbMATH DE number 1611101

    Statements

    Isomorphic embedding of \(\ell_p^n\), \(1<p<2\), into \(\ell_1^{(1+\varepsilon)n}\) (English)
    0 references
    0 references
    0 references
    0 references
    10 July 2003
    0 references
    Given two normed spaces \(X\) and \(Y\) ,we say that \(X\) \(K\)-embeds into \(Y\) provided there exists a subspace \(\overline X\) of \(Y\) and an isomorphism \(T\) from \(X\) onto \(\overline X\) with \(\|T\|\|T^{-1}\|\leq K\). In [\textit{W. B. Johnson} and \textit{G. Schechtmann}, Acta Math. 149, 71-85 (1982; Zbl 0522.46015)], it was proved that for all \(\varepsilon>0\) and \(1<p<2\), \(\ell_p^d\) \((1+\varepsilon)\)-embeds into \(\ell_1^n\) whenever \(d\leq c(p,\varepsilon)n\). The same result with \(p=2\) was previously proved in [\textit{T. Figiel, J. Lindenstrauss} and \textit{V. D. Milman}, Acta Math. 139, 53-94 (1977; Zbl 0375.52002)]. In these results, \(\varepsilon\) should be thought of as a small positive constant: The proofs in the papers cited above do not give better results (i.e., larger \(c(p,\varepsilon)\)) for \(\varepsilon>1\) than for, say, \(\varepsilon=1\). The question arises whether such better results for large \(\varepsilon\) exist. In particular, is it true that for all \(a>1\) there is a \(K=K(p,a)\) satisfying that \(\ell_p^d\) \(K\)-embeds into \(\ell_1^{ad}\)? For \(p=2\) this holds and was proved in [\textit{B. S. Kasin}, Math. USSR, Izv. 11, 317-333 (1977); translation from Izv. Akad. Nauk SSSR, Ser. Mat. 41, 334-351 (1977; Zbl 0378.46027)]. The present paper deals with the case \(1<p<2\) and the main result gives a somewhat weaker answer to the question above: For all \(a>1\) there is a \(K=K(p,a)\) satisfying that \(\ell_p^d\) \((\log d)^K\)-embeds into \(\ell_1^{ad}\). The question as posed above has recently been affirmatively solved in [\textit{W. B. Johnson} and \textit{G. Schechtman}, Geom. Funct. Anal. 13, 845-851 (2003)].
    0 references
    0 references
    0 references
    0 references
    0 references
    isomorphic embedding
    0 references
    \(\ell_p^n\) spaces
    0 references