Bounds for norms of the matrix inverse and the smallest singular value

In the first part, we obtain two easily calculable lower bounds for ‖A^-1‖, where ‖·‖ is an arbitrary matrix norm, in the case when A is an M-matrix, using first row sums and then column sums. Using those results, we obtain the characterization of M-matrices whose inverses are stochastic matrices. With different approach, we give another easily calculable lower bounds for ‖A^-1‖_∞ and ‖A^-1‖₁ in the case when A is an M-matrix. In the second part, using the results from the first part, we obtain our main result, an easily calculable upper bound for ‖A^-1‖₁ in the case when A is an SDD matrix, thus improving the known bound. All mentioned norm bounds can be used for bounding the smallest singular value of a matrix.     

Generalized induced norms

Let ‖·‖ be a norm on the algebra ?_n of all n × n matrices over ?. An interesting problem in matrix theory is that “Are there two norms ‖·‖₁ and ‖·‖₂ on ?_n such that ‖A‖ = max|‖Ax‖₂: ‖x‖₁ = 1} for all A ∈ ?_n?” We will investigate this problem and its various aspects and will discuss some conditions under which ‖·‖₁ = ‖·‖₂.     

On some subclasses of P‐matrices

A matrix with positive row sums and all its off‐diagonal elements bounded above by their corresponding row averages is called a B‐matrix by J. M. Peña in References (SIAM J. Matrix Anal. Appl. 2001; 22 :1027–1037) and (Numer. Math. 2003; 95 :337–345). In this paper, it is generalized to more extended matrices—MB‐matrices, which is proved to be a subclass of the class of P‐matrices. Subsequently, we establish relationships between defined and some already known subclasses of P‐matrices (see, References SIAM J. Matrix Anal. Appl. 2000; 21 :67–78; Linear Algebra Appl. 2004; 393 :353–364; Linear Algebra Appl. 1995; 225 :117–123). As an application, some subclasses of P‐matrices are used to localize the real eigenvalues of a real matrix. Copyright © 2007 John Wiley & Sons, Ltd.     

Perturbation bounds for the periodic Schur decomposition

This paper is devoted to the perturbation analysis for periodic regular matrix pairs. We present perturbation bounds for the periodic Schur decomposition of periodic regular matrix pairs with distinct eigenvalues, which extend the main result of Sun (SIAM J. Matrix Anal. Appl. 16:1328–1340, 1995). The results are illustrated by a numerical example.     

Stability properties of superoptimal preconditioner from numerical range

A matrix is said to be stable if the real parts of all the eigenvalues are negative. In this paper, for any matrix A_n, we give some sufficient and necessary conditions for the stability of superoptimal preconditioner E_U(A_n) proposed by Tyrtyshnikov (SIAM J. Matrix Anal. Appl. 1992; 13 :459–473). Copyright © 2005 John Wiley & Sons, Ltd.     

On the Block Independence in G-Inverse and Reflexive Inner Inverse of A Partitioned Matrix

By applying the multiple quotient singular value decomposition QQQQQ-SVD, we study the block independence in g-inverse and reflexive inner inverse of 2× 2 partitioned matrices, and prove a conjecture in [Yiju Wang, SIAM J. Matrix Anal. Appl., 19（2）, 407-415（1998）].     

Prescribing the behavior of early terminating GMRES and Arnoldi iterations

We generalize and extend results of the series of papers by Greenbaum and Strako? (IMA Vol Math Appl 60:95–118, 1994), Greenbaum et al. (SIAM J Matrix Anal Appl 17(3):465–469, 1996), Arioli et al. (BIT 38(4):636–643, 1998) and Duintjer Tebbens and Meurant (SIAM J Matrix Anal Appl 33(3):958–978, 2012). They show how to construct matrices with right-hand sides generating a prescribed GMRES residual norm convergence curve as well as prescribed Ritz values in all iterations, including the eigenvalues, and give parametrizations of the entire class of matrices and right-hand sides with these properties. These results assumed that the underlying Arnoldi orthogonalization processes are breakdown-free and hence considered non-derogatory matrices only. We extend the results with parametrizations of classes of general nonsingular matrices with right-hand sides allowing the early termination case and also give analogues for the early termination case of other results related to the theory developed in the papers mentioned above.     

Iterative minimization of the Rayleigh quotient by block steepest descent iterations

The topic of this paper is the convergence analysis of subspace gradient iterations for the simultaneous computation of a few of the smallest eigenvalues plus eigenvectors of a symmetric and positive definite matrix pair (A,M). The methods are based on subspace iterations for A ^{? 1}M and use the Rayleigh‐Ritz procedure for convergence acceleration. New sharp convergence estimates are proved by generalizing estimates, which have been presented for vectorial steepest descent iterations (see SIAM J. Matrix Anal. Appl., 32(2):443‐456, 2011). Copyright © 2013 John Wiley & Sons, Ltd.     

A block coordinate gradient descent method for regularized convex separable optimization and covariance selection

We consider a class of unconstrained nonsmooth convex optimization problems, in which the objective function is the sum of a convex smooth function on an open subset of matrices and a separable convex function on a set of matrices. This problem includes the covariance selection problem that can be expressed as an ℓ ₁-penalized maximum likelihood estimation problem. In this paper, we propose a block coordinate gradient descent method (abbreviated as BCGD) for solving this class of nonsmooth separable problems with the coordinate block chosen by a Gauss-Seidel rule. The method is simple, highly parallelizable, and suited for large-scale problems. We establish global convergence and, under a local Lipschizian error bound assumption, linear rate of convergence for this method. For the covariance selection problem, the method can terminate in O(n³/e){O(n^3/\epsilon)} iterations with an e{\epsilon}-optimal solution. We compare the performance of the BCGD method with the first-order methods proposed by Lu (SIAM J Optim 19:1807–1827, 2009; SIAM J Matrix Anal Appl 31:2000–2016, 2010) for solving the covariance selection problem on randomly generated instances. Our numerical experience suggests that the BCGD method can be efficient for large-scale covariance selection problems with constraints.     

A note on spectrum analysis of augmentation block Schur complement preconditioners

In this note, as a generalization of the preconditioner presented by Greif et al. (SIAM J Matrix Anal Appl 27:779–792, 2006), we consider a set of augmentation block Schur complement preconditioners for solving saddle point systems whose coefficient matrices have singular (1,1) blocks. The spectral properties of the preconditioned matrices are analyzed and an optimal preconditioner is derived.     

Perturbation Theory for Linearly Perturbed Algebraic Riccati Equations

The perturbation results for the solutions of two linearly perturbed algebraic Riccati equations are derived. We generalize the results of Sun [SIAM J. Matrix Anal. Appl., 19 (1998):39–65] for continuous (CARE) and discrete (DARE) algebraic Riccati equations, respectively. The results are illustrated by numerical examples.     

On an Ingelstam's Theorem

A well-known Ingelstam's Theorem asserts that every real Hilbert space A with an associative unital product satisfying ‖ xy‖ ≤ ‖ x‖ ‖ y‖ and ‖ 1‖ = 1 is isomorphic to the reals ?, or the complex numbers ?, or the quaternions ?. This note deals with a nonunital and nonassociative extension of the Ingelstam Theorem. So the assumptions about associativity and existence of unity are weakened to the existence of a nonzero central idempotent e such that ‖ ex‖ = ‖e‖ ‖ x‖ for all x, and that in A holds a determined kind of algebraic identity strictly weaker that alternativeness. We prove that, up to isomorphisms, there are only seven algebras satisfying these assumptions, even without the requirement of completeness. On the other hand, Section 3 presents another characterization of the obtained algebras with the flavor of one of the main theorems in Bhatt et al. (1998 Bhatt , S. J. , Karia , D. J. , Kulkarni , S. H. , Shimpi , M. E. ( 1998 ). A note on the Gelfand-Mazur theorem . Proc. Amer. Math. Soc. 126 ( 10 ): 2999 – 3005 .[Crossref], [Web of Science ®] , [Google Scholar]).     

PROBABILISTIC OSTROWSKI TYPE INEQUALITIES

New very general univariate and multivariate probabilistic Ostrowski type inequalities are established, involving ‖·‖_∞ and ‖·‖ _p , p≥1 norms of probability density functions. Some of these inequalities provide pointwise estimates to the error of probability distribution function from the expectation of some simple function of the engaged random variable. Other inequalities give upper bounds for the expectation and variance of a random variable. All are done over finite domains. At the end are given applications, especially for the Beta random variable.     

Efficient computation of spectral bounds for Hessian matrices on hyperrectangles for global optimization

We compare two established and a new method for the calculation of spectral bounds for Hessian matrices on hyperrectangles by applying them to a large collection of 1,522 objective and constraint functions extracted from benchmark global optimization problems. Both the tightness of the spectral bounds and the computational effort of the three methods, which apply to $C^2$ functions ${\varphi }:\mathbb{R }^n\rightarrow \mathbb{R }$ that can be written as codelists, are assessed. Specifically, we compare eigenvalue bounds obtained with the interval variant of Gershgorin’s circle criterion (Adjiman et al. in Comput Chem Eng 22(9):1137–1158, 1998; Gershgorin in Izv. Akad. Nauk SSSR, Ser. fizmat. 6:749–754, 1931), Hertz (IEEE Trans Autom Control 37:532–535, 1992) and Rohn’s (SIAM J Matrix Anal Appl 15(1):175–184, 1994) method for tight bounds of interval matrices, and a recently proposed Hessian matrix eigenvalue arithmetic (Mönnigmann in SIAM J. Matrix Anal. Appl. 32(4): 1351–1366, 2011), which deliberately avoids the computation of interval Hessians. The eigenvalue arithmetic provides tighter, as tight, and less tight bounds than the interval variant of Gershgorin’s circle criterion in about 15, 61, and 24 % of the examples, respectively. Hertz and Rohn’s method results in bounds that are always as tight as or tighter than those from Gershgorin’s circle criterion, and as tight as or tighter than those from the eigenvalue arithmetic in 96 % of the cases. In 4 % of the examples, the eigenvalue arithmetic results in tighter bounds than Hertz and Rohn’s method. This result is surprising, since Hertz and Rohn’s method provides tight bounds for interval matrices. The eigenvalue arithmetic provides tighter bounds in these cases, since it is not based on interval matrices.     

A contribution to perturbation analysis for total least squares problems

In this note, we present perturbation analysis for the total least squares (Tls) problems under the genericity condition. We review the three condition numbers proposed respectively by Zhou et al. (Numer. Algorithm, 51 (2009), pp. 381–399), Baboulin and Gratton (SIAM J. Matrix Anal. Appl. 32 (2011), pp. 685–699), Li and Jia (Linear Algebra Appl. 435 (2011), pp. 674–686). We also derive new perturbation bounds.     

On the existence of unitarily invariant norm under some conditions

Let r ₁, …, r _m be positive real numbers and A ₁, …, A _m be n × n matrices with complex entries. In this article, we present a necessary and sufficient condition for the existence of a unitarily invariant norm ‖·‖, such that ‖A _i ‖ = r _i , for i = 1, …, m. Then we identify the greatest unitarily invariant norm which satisfies this condition. Using this, we get an approximation of unitarily invariant norms. Although the minimum unitarily invariant norm which satisfies this condition does not exist in general, we find conditions over A _i s and r _i s which are sufficient for the existence of such a norm. Finally, we get a characterization of unitarily invariant norms.     

Real solutions of the equation ϕ(X)=1/nJ n

For ann x n real matrixX, let ?(X)=X ο (X ^?1) ^T , where ο stands for the Hadamard (entrywise) product. SupposeA, B, C andD aren x n real nonsingular matrices, and among them there are at least one solutions to the equation ?(X)=1/nJ _n . An equivalent condition which enable $M = \left( {\begin{array}{*{20}c} A & B \\ C & D \\ \end{array} } \right)$ become a real solution to the equation ?(X)=1/2nJ _2n , is given. As applications, we get new real solutions to the matrix equation ?(X)-1/2nJ _2n by applying the results of Zhang, Yang and Cao [SIAM. J. Matrix Anal. Appl, 21 (1999), pp: 642–645] and Chen [SIAM. J. Matrix Anal. Appl, 22 (2001), pp:965–970]. At the same time, all solutions of the matrix equation ?(X)=1/4J ₄ are also given.     

A note on computational formulas for the Drazin inverse of certain block matrices

In this paper, we give a computational formula for the Drazin inverse of the sum P+Q, then applying it we give some computational formulas for the Drazin inverse of block matrix (A and D are square) with generalized Schur complement S=D?CA ^D B is nonsingular under some conditions. These results extend the results about the Drazin inverse of M given by R. Hartwig, X. Li and Y.?Wei (SIAM J. Matrix Anal. Appl. 27:757?C771, 2006) and by C. Deng (J. Math. Anal. Appl. 368:1?C8, 2010).     

Sharp bounds for integral means

We introduce a bound M of f, ‖f‖_∞?M?2‖f‖_∞, which allows us to give for 0?p<∞ sharp upper bounds, and for −∞<p<0 sharp lower bounds for the average of |f|^p over E if the average of f over E is zero. As an application we give a new proof of Grüss's inequality estimating the covariance of two random variables. We also give a new estimate for the error term in the trapezoidal rule.     

The generalized superoptimal preconditioner

For any given n-by-n matrix A, a specific circulant preconditioner t_F(A) introduced by Tyrtyshnikov [E. Tyrtyshnikov, Optimal and super-optimal circulant preconditioners, SIAM J. Matrix Anal. Appl. 13 (1992) 459-473] is defined to be the solution of