Decomposition of a scalar matrix into a sum of orthogonal projections (Q1405039): Difference between revisions

A Hermitian matrix \(P=P^2=P^*\) is called an (orthogonal) projection. By \textit{P. A. Fillmore} [J. Funct. Anal. 4, 146-152 (1969; Zbl 0176.43404)], a Hermitian matrix \(A\) is a sum of projections if and only if tr \(A\in {\mathbb Z}\) and \(\text{tr }A\geq\text{rank} A\). The authors prove that the matrix \(\alpha I_n\) is a sum of projections if and only if \(\alpha \geq 1\) and \(\alpha n\in {\mathbb N}\). The number of projections in this sum is at most \(\alpha n\). They obtain the minimal number of summands in a decomposition of \(\alpha I_n\) into a sum of projections and for a given \(k\) they describe the set of couples \((\alpha ,n)\) for which \(\alpha I_n\) is a sum of \(k\) projections.