An algorithm for constructing doubly stochastic matrices for the inverse eigenvalue problem

Abstract

In this note, we present an algorithm that yields many new methods for constructing doubly stochastic and symmetric doubly stochastic matrices for the inverse eigenvalue problem. In addition, we introduce new open problems in this area that lay the ground for future work.(possibly complex) to be the spectrum of an n × n nonnegative matrix A. Although this inverse eigenvalue problem has attracted a considerable amount of interest, for n > 3 it is still unsolved except in restricted cases. Generally, we have two cases that result in three problems.• When λ = (λ 1 , ..., λ n ) is complex, little is known. The case n ≤ 3 was completely solved in [17] and the solution of the 4 × 4 trace-zero nonnegative inverse eigenvalue problem was given in [23].• When the n-tuples (λ 1 , . . . , λ n ) are real, then we have the following two problems: 1) The real nonnegative inverse eigenvalue problem (RNIEP) asks which sets of n real numbers occur as the spectrum of an n × n nonnegative matrix A.2) The symmetric nonnegative inverse eigenvalue problem (SNIEP) asks which sets of n real numbers can occur as the spectrum of an n × n symmetric nonnegative matrix A.Each problem remains open. Many partial results for the three problems are known see the recent book [10] and the references therein, for the collection of all known results concerning these problems. Various people raised the question whether the (RNIEP) and (SNIEP) are generally equivalent. In the low dimension n ≤ 4, the two problems are actually equivalent (see [9]). However, for n > 4, the paper [14] showed that the two problems are generally different. In addition, the paper [9] gives a construction for data λ = (λ 1 , ..., λ 5 ) which is a solution for the (RNIEP) and there is no symmetric nonnegative 5 × 5 matrix with spectrum λ.Another object of study in this area that has a big interest is the inverse eigenvalue problem for nonnegative matrices with extra properties. For example, we can consider the same problems for doubly stochastic matrices. So that we have the following problems.Problem 1.1. The doubly stochastic inverse eigenvalue problem denoted by (DIEP), is the problem of determining the necessary and sufficient conditions for a complex n-tuples to be the spectrum of an n × n doubly stochastic matrix. Equivalently, this problem can also be characterized as the problem of finding the region Θ n of C n such that any point in Θ n is the spectrum of an n × n doubly stochastic matrix. Now, when the n-tuples (λ 1 , . . . , λ n ) are all real, then we have the following two problems:

Generalization of some results concerning eigenvalues of a certain class of matrices and some applications

“…In this section, we study some of the effects of Theorem 2.3 on the spectral properties of doubly stochastic matrices which has been the object of study for a long time (see [9,13,14,15,16,20,21] and the reference therein).…”

Inverse problems for symmetric doubly stochastic matrices whose Suleĭmanova spectra are bounded below by 1/2

“…Inverse eigenvalue problems for classes of matrices such as (symmetric or not) matrices with non-negative entries, stochastic, or doubly stochastic are well-rooted in the literature, having their origin in the works of Suleȋmanova [19,20] and, independently, Perfect ( [14,15]) with an important subsequent continuation by Perfect and Mirsky [16]. Recently, the problem has gained new impetus as reflected by a plethora of new sufficient conditions ( [2,5,6,7,10,11]). We refer to Mourad's paper [9] for a good overview concerning the said problems.…”

On Spectral Properties of Doubly Stochastic Matrices

“…The inverse eigenvalue problems for non-negative doubly stochastic matrices have its origin in work of [2][3][4][5]. For more details on inverse eigenvalue problems, we refer the reader to [6][7][8][9][10] and references therein.…”

“…In the literature [6,7,10,14,15], much attention has been payed to study the eigenvalues or eigen-spectrum of doubly stochastic matrices. According to the best of our knowledge, however, the study of spectral properties, such as structured singular values for doubly stochastic matrices, is utterly missing from the literature.…”