On the doubly stochastic realization of spectra (Q2157347)

Let \(\lambda = (r;{\lambda _2},\dots,{\lambda _n})\) be a list of numbers with \(r > {\max _{2 \leqslant j \leqslant n}}\left| {{\lambda _j}} \right|\); this is said to be realizable if \(\lambda \) is the spectrum of an \(n \times n\) nonnegative matrix \(A\). In this case, \(A\) is said to be a nonnegative realization of \(\lambda \). Moreover, if each row and column sum of \(A\) is equal to \(r\), then \(\lambda \) is said to be doubly stochastically realizable and in such a case \(A\) is said to be a doubly stochastic realization of \(\lambda \). In this article, it is first introduced a method for passing from a nonnegative realization to a doubly stochastic realization. A new sufficient condition for a stochastic matrix \(A\) to be cospectral to a doubly stochastic matrix \(B\) is given and in this case \(B\) is shown to be the unique closest doubly stochastic matrix to \(A\) with respect to the Frobenius norm. Further, an improvement of \textit{W. Guo}'s bound [Linear Algebra Appl. 266, 261--270 (1997; Zbl 0903.15003)] is formulated. This sharpens the existing known bound for the case when one or at least one of \(\{ {\lambda _2},\dots,{\lambda _n}\} \) is real. In addition, the case when \(\{ {\lambda _2},\dots,{\lambda _n}\} \) are all non-real, which has not been dealt before, is analysed. Here, the main results of the paper sharpen Guo's bound and for doubly stochastic realizations an upper bound that improves Guo's bound as well is formulated. Finally, for some certain special cases, a further improvement of the last bound for doubly stochastic realization is obtained. It should be noted here that the main tool used is \textit{A. Brauer}'s theorem [Duke Math. J. 19, 75--91 (1952; Zbl 0046.01202)].