The finite dimensional basis problem with an appendix on nets of Grassmann manifolds (Q759961): Difference between revisions

The main result of the paper under consideration says that there is a constant \(C>0\) such that, for every \(n\in {\mathbb{N}}\), one can find a 2n- dimensional space B, for which each projection \(P:B\to B\) of rank n has an operator norm greater than \(C\sqrt{n}\). This is an answer in the negative to the ''finite-dimensional basis problem'', where one asks for a constant \(C>0\) such that every finite-dimensional space, denote its dimension by n, has a basis \((x_ j)\) satisfying \(\| \sum^{k}_{j=1}t_ jx_ j\| \leq \| \sum^{n}_{j=1}t_ jx_ j\|\) for all \(t_ 1,...,t_ n\in {\mathbb{R}}\) and \(1\leq k\leq n\). According to the author, \textit{E. D. Gluskin} [Dokl. Akad. Nauk SSSR 261, 1046-1050 (1981; Zbl 0501.46013)] has also obtained the latter, and \textit{G. Pisier} [Acta Math. 151, 181-208 (1983; Zbl 0542.46038)] has constructed an infinite dimensional Banach space, where each projection of rank n has an operator norm of order \(\sqrt{n}\). The proof of the main result allows an significantly stronger assertion than that stated above (operators with finite rank instead of projectors).