JAM Journal of Applied Mathematics 1687-0042 1110-757X Hindawi Publishing Corporation 513513 10.1155/2014/513513 513513 Research Article Extremal Inverse Eigenvalue Problem for a Special Kind of Matrices Liu Zhibing 1 Xu Yeying 2 Wang Kanmin 1 Xu Chengfeng 1 Shen Hui-Shen 1 School of Sciences Jiujiang University Jiujiang 332005 China jju.edu.cn 2 Faculty of Library Jiujiang University Jiujiang 332005 China jju.edu.cn 2014 522014 2014 11 06 2013 13 12 2013 5 2 2014 2014 Copyright © 2014 Zhibing Liu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

We consider the following inverse eigenvalue problem: to construct a special kind of matrix (real symmetric doubly arrow matrix) from the minimal and maximal eigenvalues of all its leading principal submatrices. The necessary and sufficient condition for the solvability of the problem is derived. Our results are constructive and they generate algorithmic procedures to construct such matrices.

1. Introduction

Peng et al. in  solved two inverse eigenvalue problems for symmetric arrow matrices and, in the other article , a correction, for one of the problems stated in , has been presented as well. In recent paper , Nazari and Beiranvand introduced an algorithm to construct symmetric quasi- antibidiagonal matrices that having its given eigenvalues. Pickmann et al. in  introduced an algorithm for inverse eigenvalue problem on symmetric tridiagonal matrices. In this paper we introduced symmetric doubly arrow matrix as follows: (1) A = ( a 1 b 1 b s - 1 0 0 b 1 a 2 0 0 0 0 0 0 b s - 1 0 a s b s b n - 1 0 0 b s a s + 1 0 0 0 0 b n - 1 0 a n ) , a j , b j , where b j 0 ,   1 s n . If s = 1 or s = n ; then the matrix A of the form (1) is a symmetric arrow matrix as follows: (2) B = ( a 1 b 1 b n - 1 b 1 a 2 0 b n - 1 0 a n ) , a j , b j .

This family of matrices appears in certain symmetric inverse eigenvalue and inverse Sturm-Liouville problems [5, 6], which arise in many applications , including modern control theory and vibration analysis [7, 8]. In this paper, we construct matrix A of the form (1), from a special kind of spectral information, which only recently is being considered. Since this type of matrix structure generalizes the well-known arrow matrices, we think that it will also become of interest in applications.

We will denote I j as the identity matrix of order j ; A j as the j × j leading principal submatrix of A ; P j ( λ ) as the characteristic polynomial of A j ; and λ 1 ( j ) λ 2 ( j ) λ j ( j ) as the eigenvalues of A j .

We want to solve the following problem.

Problem 1.

Given the 2 n - 1 real numbers λ 1 ( j ) and λ j ( j ) , j = 1,2 , , n , find an n × n matrix A of the form (1) such that λ 1 ( j ) and λ j ( j ) are, respectively, the minimal and maximal eigenvalues of A j ,   j = 1,2 , , n .

Our work is motivated by the results in . There, the authors solved this kind of inverse eigenvalue problem for symmetric arrow matrix B of the form (2).

Theorem 2 (see [<xref ref-type="bibr" rid="B11">2</xref>]).

Let the real numbers λ 1 ( j ) and λ j ( j ) , j = 1,2 , , n , be given. Then there exists an n × n matrix B of the form (2), such that λ 1 ( j ) and λ j ( j ) are, respectively, the minimal and maximal eigenvalues of its leading principal submatrix B j , j = 1,2 , , n , if and only if (3) λ 1 ( n ) λ 1 ( 3 ) λ 1 ( 2 ) λ 1 ( 1 ) λ 2 ( 2 ) λ 3 ( 3 ) λ n ( n ) .

Theorem 3 (see [<xref ref-type="bibr" rid="B11">2</xref>]).

Let the real numbers λ 1 ( j ) and λ j ( j ) ,   j = 1,2 , , n , be given. Then there exists a unique n × n matrix B of the form (2), with a j and b j > 0 , such that λ 1 ( j ) and λ j ( j ) are, respectively, the minimal and the maximal eigenvalues of its leading principal submatrix B j ,   j = 1,2 , , n , if and only if (4) λ 1 ( n ) < < λ 1 ( 3 ) < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < λ 3 ( 3 ) < < λ n ( n ) .

In this paper, we will show that Theorems 2 and 3 are also right for symmetric doubly arrow matrix A in (1) by a similar method.

The paper is organized as follows. In Section 2, we discuss some properties of A . In Section 3, we solve Problem 1 by giving a necessary and sufficient condition for the existence of the matrix A in (1) and also solve the case in which the matrix A , in Problem 1, is required to have all its entries b i positive. Finally, In Section 4 we show some examples to illustrate the results.

2. Properties of the Matrix <inline-formula> <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M56"> <mml:mrow> <mml:mi>A</mml:mi></mml:mrow> </mml:math></inline-formula> Lemma 4.

Let A be a matrix of the form (1). Then the sequence of characteristic polynomials { P j ( λ ) } j = 1 n satisfies the recurrence relation: (5) p 1 ( λ ) = ( λ - a 1 ) , p j ( λ ) = ( λ - a j ) p j - 1 ( λ ) - b j - 1 2 i = 2 j - 1 ( λ - a i ) , 222222222222222222 j = 2,3 , , s , p j ( λ ) = ( λ - a j ) p j - 1 ( λ ) - b j - 1 2 i = s + 1 j - 1 ( λ - a i ) p s - 1 ( λ ) , 22222222222222222222222 j = s + 1 , , n .

Proof.

It is easy to verify by expanding the determinant.

Lemma 5 (see [<xref ref-type="bibr" rid="B11">2</xref>]).

Let p ( λ ) be a monic polynomial of degree n with all real zeroes. If λ 1 and λ n are, respectively, the minimal and maximal zeroes of p ( λ ) , then

if μ < λ 1 , we have that ( - 1 ) n p ( μ ) > 0 ;

if μ > λ n , we have that p ( μ ) > 0 .

Observe that, from the Cauchy interlacing property, the minimal and the maximal eigenvalue, λ 1 ( j ) and λ j ( j ) , respectively, of each leading principal submatrix A j ,   j = 1,2 , , n , of the matrix A in (1) satisfy the relations (6) λ 1 ( n ) λ 1 ( 3 ) λ 1 ( 2 ) λ 1 ( 1 ) λ 2 ( 2 ) λ 3 ( 3 ) λ n ( n ) , (7) λ 1 ( j ) a i λ j ( j ) , i = 1,2 , , j ;    j = 1,2 , , n .

Lemma 6.

Let { P j ( λ ) } j = 1 n be the polynomials defined in (5), whose minimal and maximal zeroes, λ 1 ( j ) and λ j ( j ) ,   j = 1,2 , , n , respectively, satisfy the relations (6) and (7), and (8) h j = { p j - 1 ( λ 1 ( j ) ) i = 2 j - 1 ( λ j ( j ) - a i ) - p j - 1 ( λ j ( j ) ) i = 2 j - 1 ( λ 1 ( j ) - a i ) , m m m m m m m m m m m m m m m m m m j = 2,3 , , s , p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ j ( j ) ) × i = s + 1 j - 1 ( λ j ( j ) - a i ) - p j - 1 ( λ j ( j ) ) p s - 1 ( λ 1 ( j ) ) × i = s + 1 j - 1 ( λ 1 ( j ) - a i ) , j = s + 1 , , n . Then (9) h ~ j = ( - 1 ) j - 1 h j 0 , j = 2,3 , , n .

Proof.

From Lemma 5, we have (10) ( - 1 ) j - 1 p j - 1 ( λ 1 ( j ) ) 0 , p j - 1 ( λ j ( j ) ) 0 , 222222222222222222222222 j = 2,3 , , n , ( - 1 ) s - 1 p s - 1 ( λ 1 ( j ) ) 0 , p s - 1 ( λ j ( j ) ) 0 , 2222222222222222222222 j = s + 1 , , n . Moreover, from (7) (11) i = 2 j - 1 ( λ j ( j ) - a i ) 0 , ( - 1 ) j i = 2 j - 1 ( λ 1 ( j ) - a i ) 0 , 2222222222222222222222222222 j = 2,3 , , s , i = s + 1 j - 1 ( λ j ( j ) - a i ) 0 , ( - 1 ) j - s + 1 i = s + 1 j - 1 ( λ 1 ( j ) - a i ) 0 , 2222222222222222222222222222222 j = s + 1 , , n . Clearly h ~ j 0 from (10) and (11).

Lemma 7 (see [<xref ref-type="bibr" rid="B11">2</xref>]).

Let A be a matrix of the form (2) with b i 0 , i = 1,2 , , n - 1 . Let λ 1 ( j ) and λ j ( j ) , respectively, be the minimal and maximal eigenvalues of the leading principal submatrix A j ,   j = 1,2 , , n , of A . Then (12) λ 1 ( j ) < < λ 1 ( 3 ) < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < λ 3 ( 3 ) < < λ j ( j ) , λ 1 ( j ) < a i < λ j ( j ) , i = 2,3 , , j for each j = 2,3 , , n .

3. Solution of Problem <xref ref-type="statement" rid="problem1">1</xref>

The following theorem solves Problem 1. In particular, the theorem shows that condition (6) is necessary and sufficient for the existence of the matrix A in (1).

Theorem 8.

Let the real numbers λ 1 ( j ) and λ j ( j ) ,   j = 1,2 , , n , be given. Then there exists an n × n matrix A of the form (1), such that λ 1 ( j ) and λ j ( j ) are, respectively, the minimal and maximal eigenvalues of its leading principal submatrix A j ,   j = 1,2 , , n , if and only if (13) λ 1 ( n ) λ 1 ( 3 ) λ 1 ( 2 ) λ 1 ( 1 ) λ 2 ( 2 ) λ 3 ( 3 ) λ n ( n ) .

Proof.

Let λ 1 ( j ) and λ j ( j ) ,   j = 1,2 , , n , satisfy (13). Observe that (14) A 1 = ( a 1 ) = ( λ 1 ( 1 ) ) and p 1 ( λ ) = λ - a 1 . From Theorem 2, there exists A j ,   j = 2 , , s with λ 1 ( j ) and λ j ( j ) as its minimal and maximal eigenvalues, respectively. To show the existence of A j ,   j = s + 1 , , n with λ 1 ( j ) and λ j ( j ) as its minimal and maximal eigenvalues, respectively, is equivalent to showing that the system of equations (15) p j ( λ 1 ( j ) ) = ( λ 1 ( j ) - a j ) p j - 1 ( λ 1 ( j ) ) - b j - 1 2 i = s + 1 j - 1 ( λ 1 ( j ) - a i ) p s - 1 ( λ 1 ( j ) ) = 0 , p j ( λ j ( j ) ) = ( λ j ( j ) - a j ) p j - 1 ( λ j ( j ) ) - b j - 1 2 i = s + 1 j - 1 ( λ j ( j ) - a i ) p s - 1 ( λ j ( j ) ) = 0 has real solution a j and b j - 1 ,   j = s + 1 , , n . If the determinant (16) h j = p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ j ( j ) ) × i = s + 1 j - 1 ( λ j ( j ) - a i ) - p j - 1 ( λ j ( j ) ) p s - 1 ( λ 1 ( j ) ) × i = s + 1 j - 1 ( λ 1 ( j ) - a i ) of the coefficient matrix of the system (15) is nonzero, then the system has unique solutions a j and b j - 1 2 ,   j = s + 1 , , n . In this case, from Lemma 6 we have h ~ j > 0 . By solving the system (15) we obtain (17) a j = ( i = s + 1 j - 1 λ 1 ( j ) p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ j ( j ) ) × i = s + 1 j - 1 ( λ j ( j ) - a i ) - λ j ( j ) p j - 1 ( λ j ( j ) ) p s - 1 ( λ 1 ( j ) ) × i = s + 1 j - 1 ( λ 1 ( j ) - a i ) ) ( h j ) - 1 , b j - 1 2 = ( λ j ( j ) - λ 1 ( j ) ) p j - 1 ( λ 1 ( j ) ) p j - 1 ( λ j ( j ) ) h j . Since (18) ( - 1 ) j - 1 ( λ j ( j ) - λ 1 ( j ) ) p j - 1 ( λ 1 ( j ) ) p j - 1 ( λ j ( j ) ) 0 ; then b j - 1 is a real number and therefore, there exists A with the spectral properties required.

Now we will show that, if h j = 0 , the system (15) still has a solution. We do this by induction by showing that the rank of the coefficients matrix is equal to the rank of the augmented matrix.

Let j = s + 1 . If h s + 1 = 0 , then (19) h ~ s + 1 = ( - 1 ) s h s + 1 = ( - 1 ) s ( p s ( λ 1 ( s + 1 ) ) p s - 1 ( λ s + 1 ( s + 1 ) ) - p s ( λ s + 1 ( s + 1 ) ) p s - 1 ( λ 1 ( s + 1 ) ) ) = 0 , which, from Lemma 5, is equivalent to (20) p s ( λ 1 ( s + 1 ) ) p s - 1 ( λ s + 1 ( s + 1 ) ) = 0 , p s ( λ s + 1 ( s + 1 ) ) p s - 1 ( λ 1 ( s + 1 ) ) = 0 . In this case the augmented matrix is (21) ( p s ( λ 1 ( s + 1 ) ) p s - 1 ( λ 1 ( s + 1 ) ) λ 1 ( s + 1 ) p s ( λ 1 ( s + 1 ) ) p s ( λ s + 1 ( s + 1 ) ) p s - 1 ( λ s + 1 ( s + 1 ) ) λ s + 1 ( s + 1 ) p s ( λ s + 1 ( s + 1 ) ) ) , and the ranks of both matrices, the coefficient matrix and the augmented matrix, are equal. Hence A s + 1 exists.

Now we consider j s + 2 . If h j = 0 , then (22) h ~ j = ( - 1 ) j - 1 h j = ( - 1 ) j - 1 ( i = s + 1 j - 1 ( λ 1 ( j ) - a i ) p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ j ( j ) ) × i = s + 1 j - 1 ( λ j ( j ) - a i ) - p j - 1 ( λ j ( j ) ) p s - 1 ( λ 1 ( j ) ) × i = s + 1 j - 1 ( λ 1 ( j ) - a i ) ) = 0 . From Lemma 5 (23) p j - 1 ( λ 1 ( j ) ) = 0 p s - 1 ( λ j ( j ) ) i = s + 1 j - 1 ( λ j ( j ) - a i ) = 0 , p j - 1 ( λ j ( j ) ) = 0 p s - 1 ( λ 1 ( j ) ) i = s + 1 j - 1 ( λ 1 ( j ) - a i ) = 0 . Then h j = 0 leads us to the following cases:

λ 1 ( j ) = λ 1 ( j - 1 ) λ j - 1 ( j - 1 ) = λ j ( j ) ,

λ 1 ( j ) = λ 1 ( j - 1 ) p s - 1 ( λ 1 ( j ) ) i = s + 1 j - 1 ( λ 1 ( j ) - a i ) = 0 ,

p s - 1 ( λ j ( j ) ) i = s + 1 j - 1 ( λ j ( j ) - a i ) = 0 λ j - 1 ( j - 1 ) = λ j ( j ) ,

p s - 1 ( λ j ( j ) ) i = s + 1 j - 1 ( λ j ( j ) - a i ) = 0 p s - 1 ( λ 1 ( j ) ) i = s + 1 j - 1 ( λ 1 ( j ) - a i ) = 0 ,

and the augmented matrix is (24) ( p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ 1 ( j ) ) i = s + 1 j - 1 ( λ 1 ( j ) - a i ) λ 1 ( j ) p j - 1 ( λ 1 ( j ) ) p j - 1 ( λ j ( j ) ) p s - 1 ( λ j ( j ) ) i = s + 1 j - 1 ( λ j ( j ) - a i ) λ j ( j ) p j - 1 ( λ j ( j ) ) ) . By replacing conditions (i)–(iii) in (24), it is clear that the coefficients matrix and the augmented matrix have the same rank. From condition (iv), the system of (15) becomes (25) p j - 1 ( λ 1 ( j ) ) a j = λ 1 ( j ) p j - 1 ( λ 1 ( j ) ) , p j - 1 ( λ j ( j ) ) a j = λ j ( j ) p j - 1 ( λ j ( j ) ) . If p j - 1 ( λ 1 ( j ) ) 0 and p j - 1 ( λ j ( j ) ) 0 , then a j = λ 1 ( j ) = λ j ( j ) and from (13) (26) λ 1 ( j ) = λ 1 ( j - 1 ) = = λ 1 ( 1 ) = = λ j - 1 ( j - 1 ) = λ j ( j ) . Thus, p j - 1 ( λ 1 ( j ) ) = p j - 1 ( λ j ( j ) ) = 0 , which is a contradiction. Hence, under condition (iv) p j - 1 ( λ 1 ( j ) ) = 0 or p j - 1 ( λ j ( j ) ) = 0 and therefore the coefficients matrix and the augmented matrix have also the same rank. By taking b j - 1 2 0 , there exists a j × j matrix A j with the required spectral properties. The necessity comes from the Cauchy interlacing property.

We have seen in the proof of Theorem 8 that if the determinant h j of the coefficients matrix of the system (15) is nonzero, then the Problem 1 has a unique solution except for the sign of the b i entries.

Now we solve the Problem 1 in the case that the b i entries are required to be positive. We need the following lemma.

Lemma 9.

Let A be a matrix of the form (1) with b i 0 , i = 1,2 , , n - 1 . Let λ 1 ( j ) and λ j ( j ) , respectively, be the minimal and maximal eigenvalues of the leading principal submatrix A j ,   j = 1,2 , , n , of A . Then (27) λ 1 ( j ) < < λ 1 ( 3 ) < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < λ 3 ( 3 ) < < λ j ( j ) , λ 1 ( j ) < a i < λ j ( j ) , i = 2,3 , , j for each j = 2,3 , , n .

Proof.

From Lemma 7, (27) hold for j s . For j = s + 1 , we have from (5) (28) p s + 1 ( λ ) = ( λ - a s + 1 ) p s - b s 2 p s - 1 ( λ ) . As b s 0 , then from Lemma 7, we have p s + 1 ( λ 1 ( s ) ) 0 , p s + 1 ( λ s ( s ) ) 0 , and (29) λ 1 ( s + 1 ) < λ 1 ( s ) < < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < < λ s ( s ) < λ s + 1 ( s + 1 ) . If λ 1 ( s + 1 ) = a s + 1 or λ s + 1 ( s + 1 ) = a s + 1 , then (30) 0 = p s + 1 ( a s + 1 ) = - b s 2 p s - 1 ( a s + 1 ) contradicts b s 0 or (29) and from (7) we have (31) λ 1 ( s + 1 ) < a s + 1 < λ s + 1 ( s + 1 ) . Let j = s + 2 . Then from (5) (32) p s + 2 ( λ 1 ( s + 1 ) ) = - b s + 1 2 ( λ 1 ( s + 1 ) - a s + 1 ) p s - 1 ( λ 1 ( s + 1 ) ) 0 . In the same way p s + 2 ( λ s + 1 ( s + 1 ) ) 0 . Hence λ 1 ( s + 1 ) and λ s + 1 ( s + 1 ) are not zeroes of p s + 2 ( λ ) and from (6) (33) λ 1 ( s + 2 ) < λ 1 ( s + 1 ) < λ 1 ( s ) < < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < < λ s ( s ) < λ s + 1 ( s + 1 ) < λ s + 2 ( s + 2 ) . Now suppose that λ 1 ( s + 2 ) = a s + 2 . Then (34) 0 = p s + 2 ( a s + 2 ) = - b s + 1 2 ( a s + 2 - a s + 1 ) p s - 1 ( a s + 2 ) = - b s + 1 2 ( λ 1 s + 2 - a s + 1 ) p s - 1 ( a s + 2 ) contradicts the inequalities (31) and (33). The same occurs if we assume that λ s + 2 ( s + 2 ) = a s + 2 . Then from (7) and Lemma 7 we have (35) λ 1 ( s + 2 ) < a i < λ s + 2 ( s + 2 ) ,          i = 2,3 , , s + 2 . Now, suppose that (27) hold for s + 3 j n - 1 and consider (36) p j + 1 ( λ ) = ( λ - a j + 1 ) p j ( λ ) - b j 2 i = s + 1 j ( λ - a i ) p s - 1 ( λ ) . Since b j 0 and λ 1 ( j ) < a i < λ j ( j ) , i = 2,3 , , j , then i = s + 1 j ( λ 1 j - a i ) 0 and i = s + 1 j ( λ j j - a i ) 0 . Hence neither λ 1 j nor λ j j are zeroes of p j + 1 ( λ ) . Then from (6) we have (37) λ 1 ( j + 1 ) < λ 1 ( j ) < < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < < λ j ( j ) < λ j + 1 ( j + 1 ) . Finally, if λ 1 ( j + 1 ) = a j + 1 , then (38) 0 = p j + 1 ( a j + 1 ) = - b j 2 i = s + 1 j ( a j + 1 - a i ) p s - 1 ( a j + 1 ) = - b j 2 i = s + 1 j ( λ 1 j + 1 - a i ) p s - 1 ( λ 1 j + 1 ) contradicts (33). Then (39) λ 1 ( j + 1 ) < a i < λ j + 1 ( j + 1 ) , i = 2,3 , , j + 1 .

The following theorem solves Problem 1 with b j > 0 .

Theorem 10.

Let the real numbers λ 1 ( j ) and λ j ( j ) ,   j = 1,2 , , n , be given. Then there exists a unique n × n matrix A of the form (1), with a j and b j > 0 , such that λ 1 ( j ) and λ j ( j ) are, respectively, the minimal and maximal eigenvalues of its leading principal submatrix A j , j = 1,2 , , n , if and only if (40) λ 1 ( n ) < < λ 1 ( 3 ) < λ 1 ( 2 ) < λ 1 ( 1 ) < λ 2 ( 2 ) < λ 3 ( 3 ) < < λ n ( n ) .

Proof.

The proof is quite similar to the proof of Theorem 8: Let λ 1 ( j ) and λ j ( j ) , j = 2 , , n , satisfy (40). From Theorem 3, there exists A j ,   j = 2 , , s with λ 1 ( j ) and λ j ( j ) as its minimal and maximal eigenvalues, respectively. To show the existence of A j ,   j = s + 1 , , n with the required spectral properties, is equivalent to showing that the system of (15) has real solutions a j and b j - 1 , with b j - 1 > 0 , j = s + 1 , s + 2 , , n . To do this it is enough to show that the determinant of the coefficients matrix (41) h j = p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ j ( j ) ) × i = s + 1 j - 1 ( λ j ( j ) - a i ) - p j - 1 ( λ j ( j ) ) p s - 1 ( λ 1 ( j ) ) × i = s + 1 j - 1 ( λ 1 ( j ) - a i ) is nonzero.

From Lemmas 6 and 9 it follows that h ~ j = ( - 1 ) j - 1 h j > 0 . Hence h j 0 and the system (15) has real and unique solutions: (42) a j = ( i = s + 1 j - 1 λ 1 ( j ) p j - 1 ( λ 1 ( j ) ) p s - 1 ( λ j ( j ) ) × i = s + 1 j - 1 ( λ j ( j ) - a i ) - λ j ( j ) p j - 1 ( λ j ( j ) ) p s - 1 ( λ 1 ( j ) ) × i = s + 1 j - 1 ( λ 1 ( j ) - a i ) ) ( h j ) - 1 , b j - 1 2 = ( λ j ( j ) - λ 1 ( j ) ) p j - 1 ( λ 1 ( j ) ) p j - 1 ( λ j ( j ) ) h j , where (43) ( - 1 ) j - 1 ( λ j ( j ) - λ 1 ( j ) ) p j - 1 ( λ 1 ( j ) ) p j - 1 ( λ j ( j ) ) > 0 . Then it is clear that b j - 1 2 > 0 . Therefore, the b j - 1 can be chosen positive and then there exists a unique matrix A j with the required spectral properties. The necessity of the result comes from Lemma 9.

4. Examples

Now we give an algorithm to construct the solution A of Problem 1.

Algorithm.

Input a positive integer s and real numbers λ 1 ( j ) and λ j ( j ) ,   j = 1,2 , , n ;

let a 1 = λ 1 ( 1 ) .   p 1 ( λ ) = λ - a 1 ;

for j = 2 , , n , calculate p j ( λ ) according to (5);

compute a j and b j - 1 according to (17).

Example 1.

The following numbers  (44) λ 1 ( 5 ) λ 1 ( 4 ) λ 1 ( 3 ) λ 1 ( 2 ) λ 1 ( 1 ) λ 2 ( 2 ) - 11.2369 - 11.1921 - 10.9106 - 8.7760 - 6.0043 - 2.6295 λ 3 ( 3 ) λ 4 ( 4 ) λ 5 ( 5 ) 1.8532 8.4266 10.4020 satisfy the necessary and sufficient condition (40) of Theorem 10. Then the doubly arrow matrix with b i > 0 and s = 3 is (45) A = ( - 6.0043 3.0584 5.2453 3.0584 - 5.4011 5.2453 - 2.3357 4.4747 1.8880 4.4747 6.1414 1.8880 9.8361 ) . From the above 5 × 5 doubly arrow matrix A , we recompute the spectrum σ ( A j ) of its submatrix A j by MATLAB 7.0, j = 1,2 , , 5 , and get (46) σ ( A 1 ) λ 1 ( 1 ) ¯ = - 6.0043 , σ ( A 2 ) λ 1 ( 2 ) ¯ = - 8.7759 λ 2 ( 2 ) ¯ = - 2.6295 , σ ( A 3 ) λ 1 ( 3 ) ¯ = - 10.9106 λ 2 ( 3 ) ¯ = - 4.6837 λ 3 ( 3 ) ¯ = 1.8532 , σ ( A 4 ) λ 1 ( 4 ) ¯ = - 11.1921 λ 2 ( 4 ) ¯ = - 5.0836 λ 3 ( 4 ) ¯ = 0.2494 λ 4 ( 4 ) ¯ = 8.4266 , σ ( A 5 ) λ 1 ( 5 ) ¯ = - 11.2369 λ 2 ( 5 ) ¯ = - 5.1390 λ 3 ( 5 ) ¯ = 0.1384 λ 4 ( 5 ) ¯ = 8.0719 λ 5 ( 5 ) ¯ = 10.4020 .

Example 2.

We modify the previous example, that some given eigenvalues become equal to  (47) λ 1 ( 5 ) λ 1 ( 4 ) λ 1 ( 3 ) λ 1 ( 2 ) λ 1 ( 1 ) λ 2 ( 2 ) - 11.2369 - 10.9106 - 10.9106 - 8.7760 - 6.0043 - 6.0043 λ 3 ( 3 ) λ 4 ( 4 ) λ 5 ( 5 ) 1.8532 8.4266 10.4020 . These numbers satisfy the necessary and sufficient condition (13) of Theorem 8. One solution of Problem 1 with s = 3 is the matrix (48) A = ( - 6.0043 0 6.2090 0 - 8.7760 6.2090 - 3.0531 0 4.0562 0 8.4266 4.0562 8.9205 ) . Recomputing the spectrum of A , we have (49) σ ( A 1 ) λ 1 ( 1 ) ¯ = - 6.0043 , σ ( A 2 ) λ 1 ( 2 ) ¯ = - 8.7760 λ 2 ( 2 ) ¯ = - 6.0043 , σ ( A 3 ) λ 1 ( 3 ) ¯ = - 10.9106 λ 2 ( 3 ) ¯ = - 8.7760 λ 3 ( 3 ) ¯ = 1.8532 , σ ( A 4 ) λ 1 ( 4 ) ¯ = - 10.9106 λ 2 ( 4 ) ¯ = - 8.7760 λ 3 ( 4 ) ¯ = 1.8532 λ 4 ( 4 ) ¯ = 8.4266 , σ ( A 5 ) λ 1 ( 5 ) ¯ = - 11.2369 λ 2 ( 5 ) ¯ = - 8.7760 λ 3 ( 5 ) ¯ = 0.6980 λ 4 ( 5 ) ¯ = 8.4266 λ 5 ( 5 ) ¯ = 10.4020 .

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

This work is supported by the Natural Science Foundation of Jiangxi, China (nos. 20114BAB201015, 20122BAB201013, and 20132BAB201056), and Scientific and Technological Project of Jiangxi Education Office, China (no. KJLD13093).

Peng J. Hu X.-Y. Zhang L. Two inverse eigenvalue problems for a special kind of matrices Linear Algebra and Its Applications 2006 416 2-3 336 347 10.1016/j.laa.2005.11.017 MR2242733 ZBL1097.65053 Pickmann H. Egaña J. Soto R. L. Extremal inverse eigenvalue problem for bordered diagonal matrices Linear Algebra and its Applications 2007 427 2-3 256 271 10.1016/j.laa.2007.07.020 MR2351358 ZBL1144.65026 Nazari A. M. Beiranvand Z. The inverse eigenvalue problem for symmetric quasi anti-bidiagonal matrices Applied Mathematics and Computation 2011 217 23 9526 9531 10.1016/j.amc.2011.03.031 MR2811227 ZBL1230.15007 Pickmann H. Soto R. L. Egaña J. Salas M. An inverse eigenvalue problem for symmetrical tridiagonal matrices Computers & Mathematics with Applications 2007 54 5 699 708 10.1016/j.camwa.2006.12.035 MR2347933 ZBL1155.65329 Boley D. Golub G. H. A survey of matrix inverse eigenvalue problems Inverse Problems 1987 3 4 595 622 MR928047 ZBL0633.65036 Chu M. T. Golub G. H. Inverse Eigenvalue Problems: Theory, Algorithms, and Applications 2005 New York, NY, USA Oxford University Press 10.1093/acprof:oso/9780198566649.001.0001 MR2263317 Egaña J. C. Kuhl N. M. Santos L. C. An inverse eigenvalue method for frequency isolation in spring-mass systems Numerical Linear Algebra with Applications 2002 9 1 65 79 10.1002/nla.255 MR1874783 ZBL1071.65043 Gladwell G. M. L. Inverse problems in vibration Applied Mechanics Reviews 1986 39 Hald O. H. Inverse eigenvalue problems for Jacobi matrices Linear Algebra and Its Applications 1976 14 1 63 85 MR0447279 ZBL0328.15007 Hu X. Y. Zhang L. Peng Z. Y. The construction of a Jacobi matrix from its defective eigen-pair and a principal submatrix Mathematica Numerica Sinica 2000 22 3 345 354 MR1797955 ZBL1081.65511 Liao A. P. Bai Z. Z. Construction of positive definite Jacobian matrices from two eigenpairs Journal on Numerical Methods and Computer Applications 2002 23 2 131 138 MR1919727 Lu L. Ng M. K. On sufficient and necessary conditions for the Jacobi matrix inverse eigenvalue problem Numerische Mathematik 2004 98 1 167 176 10.1007/s00211-004-0525-x MR2076058 ZBL1058.65046