The retraction constant in some Banach spaces (Q1867275): Difference between revisions

The following theorem was proved in [\textit{Y. Benyamini} and \textit{Y. Sternfeld}, Proc. Am. Math. Soc. 88, 439-445 (1983; Zbl 0518.46010)]: Theorem. There exists an universal constant \(k\) such that for every infinite-dimensional Banach space \(X\) there exists a Lipschitzian retraction \(R\) from the unit ball \(B(X)\) onto the unit sphere \(S(X)\) with Lipschitz constant less than \(k\). This theorem suggests some quantitative problems. First off all it seems that no significant lower and no upper bound for the constant \(k\) that appears in this theorem is known. Given an infinite-dimensional Banach space \(X\), the retraction constant \(k_0(X)\) is defined as the infimum of such \(k\) for which there exists a retraction \(R:B(X)\to S(X)\) in \({\mathcal L}(k)\), where \({\mathcal L}(k)\) is the class of Lipschitz maps with constant \(k\). The only general result states that \(k_0(X)\geq 3\) for every space \(X\). Some estimation of \(k_0 (X)\) for some classical Banach space are available in the literature, for example, \(4\leq k_0(\ell_1)<31.64\); \(k_0(c_0)<35.18\), \(k_0(L^1[0,1])<9.43\), \(k_0(C[a,b])<23.31\). The purpose of the paper under review is to improve the known upper bounds for the retraction \(k_0(X)\) in some classical Banach spaces. In [\textit{K. Bolibok}, Nonlinear Anal. Forum 3, 13-23 (1998; MR 99j:46024)] it was proved that \(k_0(\ell_1) <31.64\). In the present paper, it is proved that \(k_0 (\ell_1)<22.45\). In the last part (the case of Hilbert spaces) it is proved that \(k_0(L^2[0,1]) <28.99\).