Almost minimal orthogonal projections

Abstract

The projection constant $\Pi(E):=\Pi(E, \ell_\infty)$ of a finite-dimensional Banach space $E\subset\ell_\infty$ is by definition the smallest norm of a linear projection of $\ell_\infty$ onto $E$. Fix $n\geq 1$ and denote by $\Pi_n$ the maximal value of $\Pi(\cdot)$ amongst $n$-dimensional real Banach spaces. We prove for every $\varepsilon >0$ that there exist an integer $d\geq 1$ and an $n$-dimensional subspace $E\subset\ell_1^d$ such that $\Pi_n \leq \Pi(E, \ell_1^d) +2 \varepsilon$ and the orthogonal projection $P\colon \ell_1^d\to E$ is almost minimal in the sense that $\lVert P \rVert \leq \Pi(E, \ell_1^d)+\varepsilon$. As a consequence of our main result, we obtain a formula relating $\Pi_n$ to smallest absolute value row-sums of orthogonal projection matrices of rank $n$.