-
A Fast Iterative Robust Principal Component Analysis Method
Authors:
Timbwaoga Aime Judicael Ouermi,
Jixian Li,
Chris R. Johnson
Abstract:
Principal Component Analysis (PCA) is widely used for dimensionality reduction and data analysis. However, PCA results are adversely affected by outliers often observed in real-world data. Existing robust PCA methods are often computationally expensive or exhibit limited robustness. In this work, we introduce a Fast Iterative Robust (FIR) PCA method by efficiently estimating the inliers center loc…
▽ More
Principal Component Analysis (PCA) is widely used for dimensionality reduction and data analysis. However, PCA results are adversely affected by outliers often observed in real-world data. Existing robust PCA methods are often computationally expensive or exhibit limited robustness. In this work, we introduce a Fast Iterative Robust (FIR) PCA method by efficiently estimating the inliers center location and covariance. Our approach leverages Incremental PCA (IPCA) to iteratively construct a subset of data points that ensures improved location and covariance estimation that effectively mitigates the influence of outliers on PCA projection. We demonstrate that our method achieves competitive accuracy and performance compared to existing robust location and covariance methods while offering improved robustness to outlier contamination. We utilize simulated and real-world datasets to evaluate and demonstrate the efficacy of our approach in identifying and preserving underlying data structures in the presence of contamination.
△ Less
Submitted 19 June, 2025;
originally announced June 2025.
-
Hyper-differential sensitivity analysis with respect to model discrepancy: Prior Distributions
Authors:
Joseph Hart,
Bart van Bloemen Waanders,
Jixian Li,
Timbwaoga A. J. Ouermi,
Chris R. Johnson
Abstract:
Hyper-differential sensitivity analysis with respect to model discrepancy was recently developed to enable uncertainty quantification for optimization problems. The approach consists of two primary steps: (i) Bayesian calibration of the discrepancy between high- and low-fidelity models, and (ii) propagating the model discrepancy uncertainty through the optimization problem. When high-fidelity mode…
▽ More
Hyper-differential sensitivity analysis with respect to model discrepancy was recently developed to enable uncertainty quantification for optimization problems. The approach consists of two primary steps: (i) Bayesian calibration of the discrepancy between high- and low-fidelity models, and (ii) propagating the model discrepancy uncertainty through the optimization problem. When high-fidelity model evaluations are limited, as is common in practice, the prior discrepancy distribution plays a crucial role in the uncertainty analysis. However, specification of this prior is challenging due to its mathematical complexity and many hyper-parameters. This article presents a novel approach to specify the prior distribution. Our approach consists of two parts: (1) an algorithmic initialization of the prior hyper-parameters that uses existing data to initialize a hyper-parameter estimate, and (2) a visualization framework to systematically explore properties of the prior and guide tuning of the hyper-parameters to ensure that the prior captures the appropriate range of uncertainty. We provide detailed mathematical analysis and a collection of numerical examples that elucidate properties of the prior that are crucial to ensure uncertainty quantification.
△ Less
Submitted 28 April, 2025;
originally announced April 2025.
-
Perron similarities and the nonnegative inverse eigenvalue problem
Authors:
Charles R. Johnson,
Pietro Paparella
Abstract:
The longstanding nonnegative inverse eigenvalue problem (NIEP) is to determine which multisets of complex numbers occur as the spectrum of an entry-wise nonnegative matrix. Although there are some well-known necessary conditions, a solution to the NIEP is far from known.
An invertible matrix is called a Perron similarity if it diagonalizes an irreducible, nonnegative matrix. Johnson and Paparell…
▽ More
The longstanding nonnegative inverse eigenvalue problem (NIEP) is to determine which multisets of complex numbers occur as the spectrum of an entry-wise nonnegative matrix. Although there are some well-known necessary conditions, a solution to the NIEP is far from known.
An invertible matrix is called a Perron similarity if it diagonalizes an irreducible, nonnegative matrix. Johnson and Paparella developed the theory of real Perron similarities. Here, we fully develop the theory of complex Perron similarities.
Each Perron similarity gives a nontrivial polyhedral cone and polytope of realizable spectra (thought of as vectors in complex Euclidean space). The extremals of these convex sets are finite in number, and their determination for each Perron similarity would solve the diagonalizable NIEP, a major portion of the entire problem. By considering Perron similarities of certain realizing matrices of Type I Karpelevich arcs, large portions of realizable spectra are generated for a given positive integer. This is demonstrated by producing a nearly complete geometrical representation of the spectra of $4 \times 4$ stochastic matrices.
Similar to the Karpelevich region, it is shown that the subset of complex Euclidean space comprising the spectra of stochastic matrices is compact and star-shaped. Extremal elements of the set are defined and shown to be on the boundary.
It is shown that the polyhedral cone and convex polytope of the discrete Fourier transform (DFT) matrix corresponds to the conical hull and convex hull of its rows, respectively. Similar results are established for multifold Kronecker products of DFT matrices and multifold Kronecker products of DFT matrices and Walsh matrices. These polytopes are of great significance with respect to the NIEP because they are extremal in the region comprising the spectra of stochastic matrices.
△ Less
Submitted 5 October, 2024; v1 submitted 11 September, 2024;
originally announced September 2024.
-
Analysis of natural cardinal ranking vectors for pairwise comparisons and the universal efficiency of the Perron geometric mean
Authors:
S. Furtado,
C. R. Johnson
Abstract:
In models using pair-wise (ratio) comparisons among alternatives, a cardinal ranking vector should be deduced from a reciprocal matrix. The right Perron eigenvector (RP) was traditionally used, though several other options have emerged. We consider some alternatives, mostly new, namely the entry-wise reciprocal of the left Perron vector (LP), the left singular vector (LS), the entry-wise reciproca…
▽ More
In models using pair-wise (ratio) comparisons among alternatives, a cardinal ranking vector should be deduced from a reciprocal matrix. The right Perron eigenvector (RP) was traditionally used, though several other options have emerged. We consider some alternatives, mostly new, namely the entry-wise reciprocal of the left Perron vector (LP), the left singular vector (LS), the entry-wise reciprocal of the right singular vector (RS), the arithmetic and geometric means of RP and LP (AP and GP), and of LS and RS (AS and GS). (The ranking vector AS was proposed by Gass and Rapcsák (2004)). All 8 of these vectors produce the natural vector in the consistent case. We compare them empirically, in terms of efficiency, for random matrices, as a function of the number of alternatives. It turns out that the vector GP is universily efficient, and this fact is proven. The vector GS performs better that the remaining 6 vectors. We show that, for reciprocal matrices obtained from consistent matrices by modifying one column and the corresponding row, all 8 vectors are efficient. Moreover, the cone generated by the columns is efficient.
△ Less
Submitted 2 September, 2024; v1 submitted 1 August, 2024;
originally announced August 2024.
-
Sufficient conditions for total positivity, compounds, and Dodgson condensation
Authors:
Shaun Fallat,
Himanshu Gupta,
Charles R. Johnson
Abstract:
A $n$-by-$n$ matrix is called totally positive ($TP$) if all its minors are positive and $TP_k$ if all of its $k$-by-$k$ submatrices are $TP$. For an arbitrary totally positive matrix or $TP_k$ matrix, we investigate if the $r$th compound ($1<r<n$) is in turn $TP$ or $TP_k$, and demonstrate a strong negative resolution in general. Focus is then shifted to Dodgson's algorithm for calculating the de…
▽ More
A $n$-by-$n$ matrix is called totally positive ($TP$) if all its minors are positive and $TP_k$ if all of its $k$-by-$k$ submatrices are $TP$. For an arbitrary totally positive matrix or $TP_k$ matrix, we investigate if the $r$th compound ($1<r<n$) is in turn $TP$ or $TP_k$, and demonstrate a strong negative resolution in general. Focus is then shifted to Dodgson's algorithm for calculating the determinant of a generic matrix, and we analyze whether the associated condensed matrices are possibly totally positive or $TP_k$. We also show that all condensed matrices associated with a $TP$ Hankel matrix are $TP$.
△ Less
Submitted 9 May, 2024;
originally announced May 2024.
-
Indices of diagonalizable and universal realizability of spectra
Authors:
Charles R. Johnson,
Ana I. Julio,
Ricardo L. Soto
Abstract:
A list $Λ=\{λ_{1},\ldots ,λ_{n}\}$ of complex numbers (repeats allowed) is said to be \textit{realizable} if it is the spectrum of an entrywise nonnegative matrix $A$. $Λ$ is \textit{diagonalizably realizable} if the realizing matrix $A$ is diagonalizable. $Λ$ is said to be \textit{universally realizable} if it is \textit{\ realizable} for each possible Jordan canonical form allowed by $Λ.$ Here,…
▽ More
A list $Λ=\{λ_{1},\ldots ,λ_{n}\}$ of complex numbers (repeats allowed) is said to be \textit{realizable} if it is the spectrum of an entrywise nonnegative matrix $A$. $Λ$ is \textit{diagonalizably realizable} if the realizing matrix $A$ is diagonalizable. $Λ$ is said to be \textit{universally realizable} if it is \textit{\ realizable} for each possible Jordan canonical form allowed by $Λ.$ Here, we study the connection between diagonalizable realizability and universal realizability of spectra. In particular, we establish \textit{\ indices of realizability} for diagonalizable and universal realizability. We also define the merge of two spectra and we prove a result that allow us to easily decide, in many cases, about the universal realizability of spectra.
△ Less
Submitted 14 October, 2023; v1 submitted 11 January, 2023;
originally announced January 2023.
-
Estimating and using information in inverse problems
Authors:
Wolfgang Bangerth,
Chris R. Johnson,
Dennis K. Njeru,
Bart van Bloemen Waanders
Abstract:
In inverse problems, one attempts to infer spatially variable functions from indirect measurements of a system. To practitioners of inverse problems, the concept of "information" is familiar when discussing key questions such as which parts of the function can be inferred accurately and which cannot. For example, it is generally understood that we can identify system parameters accurately only clo…
▽ More
In inverse problems, one attempts to infer spatially variable functions from indirect measurements of a system. To practitioners of inverse problems, the concept of "information" is familiar when discussing key questions such as which parts of the function can be inferred accurately and which cannot. For example, it is generally understood that we can identify system parameters accurately only close to detectors, or along ray paths between sources and detectors, because we have "the most information" for these places. Although referenced in many publications, the "information" that is invoked in such contexts is not a well understood and clearly defined quantity.
Herein, we present a definition of information density that is based on the variance of coefficients as derived from a Bayesian reformulation of the inverse problem. We then discuss three areas in which this information density can be useful in practical algorithms for the solution of inverse problems, and illustrate the usefulness in one of these areas -- how to choose the discretization mesh for the function to be reconstructed -- using numerical experiments.
△ Less
Submitted 10 February, 2025; v1 submitted 18 August, 2022;
originally announced August 2022.
-
$k$-NIM trees: Characterization and Enumeration
Authors:
Charles R. Johnson,
George Tsoukalas,
Greyson C. Wesley,
Zachary Zhao
Abstract:
Among those real symmetric matrices whose graph is a given tree $T$, the maximum multiplicity $M(T)$ that can be attained by an eigenvalue is known to be the path cover number of $T$. We say that a tree is $k$-NIM if, whenever an eigenvalue attains a multiplicity of $k-1$ less than the maximum multiplicity, all other multiplicities are $1$. $1$-NIM trees are known as NIM trees, and a characterizat…
▽ More
Among those real symmetric matrices whose graph is a given tree $T$, the maximum multiplicity $M(T)$ that can be attained by an eigenvalue is known to be the path cover number of $T$. We say that a tree is $k$-NIM if, whenever an eigenvalue attains a multiplicity of $k-1$ less than the maximum multiplicity, all other multiplicities are $1$. $1$-NIM trees are known as NIM trees, and a characterization for NIM trees is already known. Here we provide a graph-theoretic characterization for $k$-NIM trees for each $k\geq 1$, as well as count them. It follows from the characterization that $k$-NIM trees exist on $n$ vertices only when $k=1,2,3$. In case $k=3$, the only $3$-NIM trees are simple stars.
△ Less
Submitted 11 August, 2022; v1 submitted 10 August, 2022;
originally announced August 2022.
-
Matricial Proofs of Some Classical Results about Critical Point Location
Authors:
Charles R. Johnson,
Pietro Paparella
Abstract:
The Gauss--Lucas and Bôcher--Grace--Marden theorems are classical results in the geometry of polynomials. Proofs of the these results are available in the literature, but the approaches are seemingly different. In this work, we show that these theorems can be proven in a unified theoretical framework utilizing matrix analysis (in particular, using the field of values and the differentiator of a ma…
▽ More
The Gauss--Lucas and Bôcher--Grace--Marden theorems are classical results in the geometry of polynomials. Proofs of the these results are available in the literature, but the approaches are seemingly different. In this work, we show that these theorems can be proven in a unified theoretical framework utilizing matrix analysis (in particular, using the field of values and the differentiator of a matrix). In addition, we provide a useful variant of a well-known result due to Siebeck.
△ Less
Submitted 21 December, 2020;
originally announced December 2020.
-
Spectra of Convex Hulls of Matrix Groups
Authors:
Eric Jankowski,
Charles R. Johnson,
Derek Lim
Abstract:
The still-unsolved problem of determining the set of eigenvalues realized by $n$-by-$n$ doubly stochastic matrices, those matrices with row sums and column sums equal to $1$, has attracted much attention in the last century. This problem is somewhat algebraic in nature, due to a result of Birkhoff demonstrating that the set of doubly stochastic matrices is the convex hull of the permutation matric…
▽ More
The still-unsolved problem of determining the set of eigenvalues realized by $n$-by-$n$ doubly stochastic matrices, those matrices with row sums and column sums equal to $1$, has attracted much attention in the last century. This problem is somewhat algebraic in nature, due to a result of Birkhoff demonstrating that the set of doubly stochastic matrices is the convex hull of the permutation matrices. Here we are interested in a general matrix group $G \subseteq GL_n(\mathbb{C})$ and the hull spectrum $\text{HS}(G)$ of eigenvalues realized by convex combinations of elements of $G$. We show that hull spectra of matrix groups share many nice properties. Moreover, we give bounds on the hull spectra of matrix groups, determine $\text{HS}(G)$ exactly for important classes of matrix groups, and study the hull spectra of representations of abstract groups.
△ Less
Submitted 27 December, 2019; v1 submitted 23 September, 2019;
originally announced September 2019.
-
Eigenvalue Paths Arising From Matrix Paths
Authors:
Eric Jankowski,
Charles R. Johnson
Abstract:
It is known (see e.g. [2], [4], [5], [6]) that continuous variations in the entries of a complex square matrix induce continuous variations in its eigenvalues. If such a variation arises from one real parameter $α\in [0, 1]$, then the eigenvalues follow continuous paths in the complex plane as $α$ shifts from $0$ to $1$. The intent here is to study the nature of these eigenpaths, including their b…
▽ More
It is known (see e.g. [2], [4], [5], [6]) that continuous variations in the entries of a complex square matrix induce continuous variations in its eigenvalues. If such a variation arises from one real parameter $α\in [0, 1]$, then the eigenvalues follow continuous paths in the complex plane as $α$ shifts from $0$ to $1$. The intent here is to study the nature of these eigenpaths, including their behavior under small perturbations of the matrix variations, as well as the resulting eigenpairings of the matrices that occur at $α = 0$ and $α = 1$. We also give analogs of our results in the setting of monic polynomials.
△ Less
Submitted 23 September, 2019;
originally announced September 2019.
-
The Doubly Stochastic Single Eigenvalue Problem: A Computational Approach
Authors:
Amit Harlev,
Charles R. Johnson,
Derek Lim
Abstract:
The problem of determining $DS_n$, the complex numbers that occur as an eigenvalue of an $n$-by-$n$ doubly stochastic matrix, has been a target of study for some time. The Perfect-Mirsky region, $PM_n$, is contained in $DS_n$, and is known to be exactly $DS_n$ for $n \leq 4$, but strictly contained within $DS_n$ for $n = 5$. Here, we present a Boundary Conjecture that asserts that the boundary of…
▽ More
The problem of determining $DS_n$, the complex numbers that occur as an eigenvalue of an $n$-by-$n$ doubly stochastic matrix, has been a target of study for some time. The Perfect-Mirsky region, $PM_n$, is contained in $DS_n$, and is known to be exactly $DS_n$ for $n \leq 4$, but strictly contained within $DS_n$ for $n = 5$. Here, we present a Boundary Conjecture that asserts that the boundary of $DS_n$ is achieved by eigenvalues of convex combinations of pairs of (or single) permutation matrices. We present a method to efficiently compute a portion of $DS_n$, and obtain computational results that support the Boundary Conjecture. We also give evidence that $DS_n$ is equal to $PM_n$ for certain $n > 5$.
△ Less
Submitted 3 April, 2020; v1 submitted 9 August, 2019;
originally announced August 2019.
-
The Inverse Eigenvalue Problem for Linear Trees
Authors:
Tanay Wakhare,
Charles R. Johnson
Abstract:
We prove the sufficiency of the Linear Superposition Principle for linear trees, which characterizes the spectra achievable by a real symmetric matrix whose underlying graph is a linear tree. The necessity was previously proven in 2014. This is the most general class of trees for which the inverse eigenvalue problem has been solved. We explore many consequences, including the Degree Conjecture for…
▽ More
We prove the sufficiency of the Linear Superposition Principle for linear trees, which characterizes the spectra achievable by a real symmetric matrix whose underlying graph is a linear tree. The necessity was previously proven in 2014. This is the most general class of trees for which the inverse eigenvalue problem has been solved. We explore many consequences, including the Degree Conjecture for possible spectra, upper bounds for the minimum number of eigenvalues of multiplicity $1$, and the equality of the diameter of a linear tree and its minimum number of distinct eigenvalues, etc.
△ Less
Submitted 29 March, 2022; v1 submitted 14 June, 2019;
originally announced June 2019.
-
The Proportion of Trees that are Linear
Authors:
Tanay Wakhare,
Eric Wityk,
Charles R. Johnson
Abstract:
We study several enumeration problems connected to linear trees, a broad class which includes stars, paths, generalized stars, and caterpillars. We provide generating functions for counting the number of linear trees on $n$ vertices, characterize the asymptotic growth rate of the number of nonisomorphic linear trees, and show that the distribution of $k$-linear trees on $n$ vertices follows a cent…
▽ More
We study several enumeration problems connected to linear trees, a broad class which includes stars, paths, generalized stars, and caterpillars. We provide generating functions for counting the number of linear trees on $n$ vertices, characterize the asymptotic growth rate of the number of nonisomorphic linear trees, and show that the distribution of $k$-linear trees on $n$ vertices follows a central limit theorem.
△ Less
Submitted 19 March, 2020; v1 submitted 24 January, 2019;
originally announced January 2019.
-
Spectra of Tridiagonal Matrices over a Field
Authors:
R. S. Costas-Santos,
C. R. Johnson
Abstract:
We consider spectra of $n$-by-$n$ irreducible tridiagonal matrices over a field and of their $n-1$-by-$n-1$ trailing principal submatrices. The real symmetric and complex Hermitian cases have been fully understood: it is necessary and sufficient that the necessarily real eigenvalues are distinct and those of the principal submatrix strictly interlace. So this case is very restrictive.
By contras…
▽ More
We consider spectra of $n$-by-$n$ irreducible tridiagonal matrices over a field and of their $n-1$-by-$n-1$ trailing principal submatrices. The real symmetric and complex Hermitian cases have been fully understood: it is necessary and sufficient that the necessarily real eigenvalues are distinct and those of the principal submatrix strictly interlace. So this case is very restrictive.
By contrast, for a general field, the requirements on the two spectra are much less restrictive. In particular, in the real or complex case, the $n$-by-$n$ characteristic polynomial is arbitrary (so that the algebraic multiplicities may be anything in place of all 1's in the classical cases) and that of the principal submatrix is the complement of a lower dimensional algebraic set (and so relatively free). Explicit conditions are given.
△ Less
Submitted 23 July, 2018;
originally announced July 2018.
-
The NIEP
Authors:
Charles R. Johnson,
Carlos Marijuán,
Pietro Paparella,
Miriam Pisonero
Abstract:
The nonnegative inverse eigenvalue problem (NIEP) asks which lists of $n$ complex numbers (counting multiplicity) occur as the eigenvalues of some $n$-by-$n$ entry-wise nonnegative matrix. The NIEP has a long history and is a known hard (perhaps the hardest in matrix analysis?) and sought after problem. Thus, there are many subproblems and relevant results in a variety of directions. We survey mos…
▽ More
The nonnegative inverse eigenvalue problem (NIEP) asks which lists of $n$ complex numbers (counting multiplicity) occur as the eigenvalues of some $n$-by-$n$ entry-wise nonnegative matrix. The NIEP has a long history and is a known hard (perhaps the hardest in matrix analysis?) and sought after problem. Thus, there are many subproblems and relevant results in a variety of directions. We survey most work on the problem and its several variants, with an emphasis on recent results, and include 130 references. The survey is divided into: a) the single eigenvalue problems; b) necessary conditions; c) low dimensional results; d) sufficient conditions; e) appending 0's to achieve realizability; f) the graph NIEP's; g) Perron similarities; and h) the relevance of Jordan structure.
△ Less
Submitted 1 August, 2017; v1 submitted 31 March, 2017;
originally announced March 2017.
-
Total positivity of sums, Hadamard products and Hadamard powers: Results and counterexamples
Authors:
Shaun Fallat,
Charles R. Johnson,
Alan D. Sokal
Abstract:
We show that, for Hankel matrices, total nonnegativity (resp. total positivity) of order r is preserved by sum, Hadamard product, and Hadamard power with real exponent t \ge r-2. We give examples to show that our results are sharp relative to matrix size and structure (general, symmetric or Hankel). Some of these examples also resolve the Hadamard critical-exponent problem for totally positive and…
▽ More
We show that, for Hankel matrices, total nonnegativity (resp. total positivity) of order r is preserved by sum, Hadamard product, and Hadamard power with real exponent t \ge r-2. We give examples to show that our results are sharp relative to matrix size and structure (general, symmetric or Hankel). Some of these examples also resolve the Hadamard critical-exponent problem for totally positive and totally nonnegative matrices.
△ Less
Submitted 29 January, 2021; v1 submitted 7 December, 2016;
originally announced December 2016.
-
A matricial view of the Karpelevič Theorem
Authors:
Charles R. Johnson,
Pietro Paparella
Abstract:
The question of the exact region in the complex plane of the possible single eigenvalues of all $n$-by-$n$ stochastic matrices was raised by Kolmogorov in 1937 and settled by Karpelevič in 1951 after a partial result by Dmitriev and Dynkin in 1946. The Karpelevič result is unwieldy, but a simplification was given by Đoković in 1990 and Ito in 1997. The Karpelevič region is determined by a set of b…
▽ More
The question of the exact region in the complex plane of the possible single eigenvalues of all $n$-by-$n$ stochastic matrices was raised by Kolmogorov in 1937 and settled by Karpelevič in 1951 after a partial result by Dmitriev and Dynkin in 1946. The Karpelevič result is unwieldy, but a simplification was given by Đoković in 1990 and Ito in 1997. The Karpelevič region is determined by a set of boundary arcs each connecting consecutive roots of unity of order less than $n$. It is shown here that each of these arcs is realized by a single, somewhat simple, parametrized stochastic matrix. Other observations are made about the nature of the arcs and several further questions are raised. The doubly stochastic analog of the Karpelevič region remains open, but a conjecture about it is amplified.
△ Less
Submitted 1 December, 2016; v1 submitted 21 November, 2016;
originally announced November 2016.
-
Row Cones, Perron Similarities, and Nonnegative Spectra
Authors:
C. R. Johnson,
Pietro Paparella
Abstract:
In further pursuit of the diagonalizable \emph{real nonnegative inverse eigenvalue problem} (RNIEP), we study the relationship between the \emph{row cone} $\mathcal{C}_r(S)$ and the \emph{spectracone} $\mathcal{C}(S)$ of a Perron similarity $S$. In the process, a new kind of matrix, \emph{row Hadamard conic} (RHC), is defined and related to the D-RNIEP. Characterizations are given when…
▽ More
In further pursuit of the diagonalizable \emph{real nonnegative inverse eigenvalue problem} (RNIEP), we study the relationship between the \emph{row cone} $\mathcal{C}_r(S)$ and the \emph{spectracone} $\mathcal{C}(S)$ of a Perron similarity $S$. In the process, a new kind of matrix, \emph{row Hadamard conic} (RHC), is defined and related to the D-RNIEP. Characterizations are given when $\mathcal{C}_r(S) = \mathcal{C}(S)$, and explicit examples are given for all possible set-theoretic relationships between the two cones. The symmetric NIEP is the special case of the D-RNIEP in which the Perron similarity $S$ is also orthogonal.
△ Less
Submitted 8 November, 2016;
originally announced November 2016.
-
Perron Spectratopes and the Real Nonnegative Inverse Eigenvalue Problem
Authors:
Charles R. Johnson,
Pietro Paparella
Abstract:
Call an $n$-by-$n$ invertible matrix $S$ a \emph{Perron similarity} if there is a real non-scalar diagonal matrix $D$ such that $S D S^{-1}$ is entrywise nonnegative. We give two characterizations of Perron similarities and study the polyhedra $\mathcal{C}(S) := \{ x \in \mathbb{R}^n: S D_x S^{-1} \geq 0,~D_x := \text{diag}(x) \}$ and $\mathcal{P})(S) := \{x \in \mathcal{C}(S) : x_1 = 1 \}$, which…
▽ More
Call an $n$-by-$n$ invertible matrix $S$ a \emph{Perron similarity} if there is a real non-scalar diagonal matrix $D$ such that $S D S^{-1}$ is entrywise nonnegative. We give two characterizations of Perron similarities and study the polyhedra $\mathcal{C}(S) := \{ x \in \mathbb{R}^n: S D_x S^{-1} \geq 0,~D_x := \text{diag}(x) \}$ and $\mathcal{P})(S) := \{x \in \mathcal{C}(S) : x_1 = 1 \}$, which we call the \emph{Perron spectracone} and \emph{Perron spectratope}, respectively. The set of all normalized real spectra of diagonalizable nonnegative matrices may be covered by Perron spectratopes, so that enumerating them is of interest.
The Perron spectracone and spectratope of Hadamard matrices are of particular interest and tend to have large volume. For the canonical Hadamard matrix (as well as other matrices), the Perron spectratope coincides with the convex hull of its rows.
In addition, we provide a constructive version of a result due to Fiedler (\cite[Theorem 2.4]{f1974}) for Hadamard orders, and a constructive version of \cite[Theorem 5.1]{bh1991} for Suleĭmanova spectra.
△ Less
Submitted 19 November, 2015; v1 submitted 29 August, 2015;
originally announced August 2015.
-
Equal Entries in Totally Positive Matrices
Authors:
Miriam Farber,
Mitchell Faulk,
Charles R. Johnson,
Evan Marzion
Abstract:
We show that the maximal number of equal entries in a totally positive (resp. totally nonsingular) $n\textrm{-by-}n$ matrix is $Θ(n^{4/3})$ (resp. $Θ(n^{3/2}$)). Relationships with point-line incidences in the plane, Bruhat order of permutations, and $TP$ completability are also presented. We also examine the number and positionings of equal $2\textrm{-by-}2$ minors in a $2\textrm{-by-}n$ $TP$ mat…
▽ More
We show that the maximal number of equal entries in a totally positive (resp. totally nonsingular) $n\textrm{-by-}n$ matrix is $Θ(n^{4/3})$ (resp. $Θ(n^{3/2}$)). Relationships with point-line incidences in the plane, Bruhat order of permutations, and $TP$ completability are also presented. We also examine the number and positionings of equal $2\textrm{-by-}2$ minors in a $2\textrm{-by-}n$ $TP$ matrix, and give a relationship between the location of equal $2\textrm{-by-}2$ minors and outerplanar graphs.
△ Less
Submitted 17 September, 2013;
originally announced September 2013.
-
Continuity properties of vectors realizing points in the classical field of values
Authors:
Dan Corey,
Charles R. Johnson,
Ryan Kirk,
Brian Lins,
Ilya Spitkovsky
Abstract:
For an $n$-by-$n$ matrix $A$, let $f_A$ be its "field of values generating function" defined as $f_A\colon x\mapsto x^*Ax$. We consider two natural versions of the continuity, which we call strong and weak, of $f_A^{-1}$ (which is of course multi-valued) on the field of values $F(A)$. The strong continuity holds, in particular, on the interior of $F(A)$, and at such points $z \in \partial F(A)$ wh…
▽ More
For an $n$-by-$n$ matrix $A$, let $f_A$ be its "field of values generating function" defined as $f_A\colon x\mapsto x^*Ax$. We consider two natural versions of the continuity, which we call strong and weak, of $f_A^{-1}$ (which is of course multi-valued) on the field of values $F(A)$. The strong continuity holds, in particular, on the interior of $F(A)$, and at such points $z \in \partial F(A)$ which are either corner points, belong to the relative interior of flat portions of $\partial F(A)$, or whose preimage under $f_A$ is contained in a one-dimensional set. Consequently, $f_A^{-1}$ is continuous in this sense on the whole $F(A)$ for all normal, 2-by-2, and unitarily irreducible 3-by-3 matrices. Nevertheless, we show by example that the strong continuity of $f_A^{-1}$ fails at certain points of $\partial F(A)$ for some (unitarily reducible) 3-by-3 and (unitarily irreducible) 4-by-4 matrices. The weak continuity, in its turn, fails for some unitarily reducible 4-by-4 and untiarily irreducible 6-by-6 matrices.
△ Less
Submitted 18 July, 2013;
originally announced July 2013.
-
Solution Theory for Systems of Bilinear Equations
Authors:
Charles R. Johnson,
Helena Šmigoc,
Dian Yang
Abstract:
Bilinear systems of equations are defined, motivated and analyzed for solvability. Elementary structure is mentioned and it is shown that all solutions may be obtained as rank one completions of a linear matrix polynomial derived from elementary operations. This idea is used to identify bilinear systems that are solvable for all right hand sides and to understand solvability when the number of equ…
▽ More
Bilinear systems of equations are defined, motivated and analyzed for solvability. Elementary structure is mentioned and it is shown that all solutions may be obtained as rank one completions of a linear matrix polynomial derived from elementary operations. This idea is used to identify bilinear systems that are solvable for all right hand sides and to understand solvability when the number of equations is large or small.
△ Less
Submitted 6 February, 2013;
originally announced March 2013.
-
Nonpositive Eigenvalues of the Adjacency Matrix and Lower Bounds for Laplacian Eigenvalues
Authors:
Zachary B. Charles,
Miriam Farber,
Charles R. Johnson,
Lee Kennedy-Shaffer
Abstract:
Let $NPO(k)$ be the smallest number $n$ such that the adjacency matrix of any undirected graph with $n$ vertices or more has at least $k$ nonpositive eigenvalues. We show that $NPO(k)$ is well-defined and prove that the values of $NPO(k)$ for $k=1,2,3,4,5$ are $1,3,6,10,16$ respectively. In addition, we prove that for all $k \geq 5$, $R(k,k+1) \ge NPO(k) > T_k$, in which $R(k,k+1)$ is the Ramsey n…
▽ More
Let $NPO(k)$ be the smallest number $n$ such that the adjacency matrix of any undirected graph with $n$ vertices or more has at least $k$ nonpositive eigenvalues. We show that $NPO(k)$ is well-defined and prove that the values of $NPO(k)$ for $k=1,2,3,4,5$ are $1,3,6,10,16$ respectively. In addition, we prove that for all $k \geq 5$, $R(k,k+1) \ge NPO(k) > T_k$, in which $R(k,k+1)$ is the Ramsey number for $k$ and $k+1$, and $T_k$ is the $k^{th}$ triangular number. This implies new lower bounds for eigenvalues of Laplacian matrices: the $k$-th largest eigenvalue is bounded from below by the $NPO(k)$-th largest degree, which generalizes some prior results.
△ Less
Submitted 26 May, 2012; v1 submitted 24 August, 2011;
originally announced August 2011.
-
The critical exponent for continuous conventional powers of doubly nonnegative matrices
Authors:
Charles R. Johnson,
Brian Lins,
Olivia Walch
Abstract:
We prove that there exists an exponent beyond which all continuous conventional powers of n-by-n doubly nonnegative matrices are doubly nonnegative. We show that this critical exponent cannot be less than $n-2$ and we conjecture that it is always $n-2$ (as it is with Hadamard powering). We prove this conjecture when $n<6$ and in certain other special cases. We establish a quadratic bound for the c…
▽ More
We prove that there exists an exponent beyond which all continuous conventional powers of n-by-n doubly nonnegative matrices are doubly nonnegative. We show that this critical exponent cannot be less than $n-2$ and we conjecture that it is always $n-2$ (as it is with Hadamard powering). We prove this conjecture when $n<6$ and in certain other special cases. We establish a quadratic bound for the critical exponent in general.
△ Less
Submitted 20 August, 2010;
originally announced August 2010.
-
Matrices Totally Positive Relative to a Tree, II
Authors:
R. S. Costas-Santos,
C. R. Johnson
Abstract:
In this paper we prove that for a general tree $T$, if $A$ is T-TP, all the submatrices of $A$ associated with the deletion of pendant vertices are $P$-matrices, and $\det A>0$, then the smallest eigenvalue has an eigenvector signed according to $T$.
In this paper we prove that for a general tree $T$, if $A$ is T-TP, all the submatrices of $A$ associated with the deletion of pendant vertices are $P$-matrices, and $\det A>0$, then the smallest eigenvalue has an eigenvector signed according to $T$.
△ Less
Submitted 17 February, 2015; v1 submitted 24 March, 2010;
originally announced March 2010.
-
Convergence of Polynomial Ergodic Averages of Several Variables for some Commuting Transformations
Authors:
Michael C. R. Johnson
Abstract:
Let $(X,\mathcal{B},μ)$ be a probability space and let $T_1,..., T_l$ be $l$ commuting invertible measure preserving transformations \linebreak of $X$. We show that if $T_1^{c_1} ... T_l^{c_l}$ is ergodic for each $(c_1,...,c_l)\neq (0,...,0)$, then the averages $\frac{1}{|Φ_N|}\sum_{u\inΦ_N}\prod_{i=1}^r T_1^{p_{i1}(u)}... T_l^{p_{il}(u)}f_i$ converge in $L^2(μ)$ for all polynomials…
▽ More
Let $(X,\mathcal{B},μ)$ be a probability space and let $T_1,..., T_l$ be $l$ commuting invertible measure preserving transformations \linebreak of $X$. We show that if $T_1^{c_1} ... T_l^{c_l}$ is ergodic for each $(c_1,...,c_l)\neq (0,...,0)$, then the averages $\frac{1}{|Φ_N|}\sum_{u\inΦ_N}\prod_{i=1}^r T_1^{p_{i1}(u)}... T_l^{p_{il}(u)}f_i$ converge in $L^2(μ)$ for all polynomials $p_{ij}\colon \mathbb{Z}^d\to\mathbb{Z}$, all $f_i\in L^{\infty}(μ)$, and all Følner sequences $\{Φ_N\}_{N=1}^{\infty}$ in $\mathbb{Z}^d$.
△ Less
Submitted 17 June, 2009;
originally announced June 2009.
-
Bounded Ratios of Products of Principal Minors of Positive Definite Matrices
Authors:
H. Tracy Hall,
Charles R. Johnson
Abstract:
Considered is the multiplicative semigroup of ratios of products of principal minors bounded over all positive definite matrices. A long history of literature identifies various elements of this semigroup, all of which lie in a sub-semigroup generated by Hadamard-Fischer inequalities. Via cone-theoretic techniques and the patterns of nullity among positive semidefinite matrices, a semigroup cont…
▽ More
Considered is the multiplicative semigroup of ratios of products of principal minors bounded over all positive definite matrices. A long history of literature identifies various elements of this semigroup, all of which lie in a sub-semigroup generated by Hadamard-Fischer inequalities. Via cone-theoretic techniques and the patterns of nullity among positive semidefinite matrices, a semigroup containing all bounded ratios is given. This allows the complete determination of the semigroup of bounded ratios for 4-by-4 positive definite matrices, whose 46 generators include ratios not implied by Hadamard-Fischer and ratios not bounded by 1. For n > 4 it is shown that the containment of semigroups is strict, but a generalization of nullity patterns, of which one example is given, is conjectured to provide a finite determination of all bounded ratios.
△ Less
Submitted 16 June, 2008;
originally announced June 2008.
-
Matrices Totally Positive Relative to a Tree
Authors:
Charles R. Johnson,
Roberto S. Costas-Santos,
Boris Tadchiev
Abstract:
It is known that for a totally positive (TP) matrix, the eigenvalues are positive and distinct and the eigenvector associated with the smallest eigenvalue is totally nonzero and has an alternating sign pattern. Here, a certain weakening of the TP hypothesis is shown to yield a similar conclusion.
It is known that for a totally positive (TP) matrix, the eigenvalues are positive and distinct and the eigenvector associated with the smallest eigenvalue is totally nonzero and has an alternating sign pattern. Here, a certain weakening of the TP hypothesis is shown to yield a similar conclusion.
△ Less
Submitted 1 November, 2007; v1 submitted 6 October, 2007;
originally announced October 2007.
-
On the Positivity of the Coefficients of a Certain Polynomial Defined by Two Positive Definite Matrices
Authors:
Christopher J. Hillar,
Charles R. Johnson
Abstract:
It is shown that the polynomial \[p(t) = \text{Tr}[(A+tB)^m]\] has positive coefficients when $m = 6$ and $A$ and $B$ are any two 3-by-3 complex Hermitian positive definite matrices. This case is the first that is not covered by prior, general results. This problem arises from a conjecture raised by Bessis, Moussa and Villani in connection with a long-standing problem in theoretical physics. The…
▽ More
It is shown that the polynomial \[p(t) = \text{Tr}[(A+tB)^m]\] has positive coefficients when $m = 6$ and $A$ and $B$ are any two 3-by-3 complex Hermitian positive definite matrices. This case is the first that is not covered by prior, general results. This problem arises from a conjecture raised by Bessis, Moussa and Villani in connection with a long-standing problem in theoretical physics. The full conjecture, as shown recently by Lieb and Seiringer, is equivalent to $p(t)$ having positive coefficients for any $m$ and any two $n$-by-$n$ positive definite matrices. We show that, generally, the question in the real case reduces to that of singular $A$ and $B$, and this is a key part of our proof.
△ Less
Submitted 4 July, 2007;
originally announced July 2007.
-
The Graphs for which the Maximum Multiplicity of an Eigenvalue is Two
Authors:
Charles R. Johnson,
Raphael Loewy,
Paul Anthony Smith
Abstract:
Characterized are all simple undirected graphs $G$ such that any real symmetric matrix that has graph $G$ has no eigenvalues of multiplicity more than 2. All such graphs are partial 2-trees (and this follows from a result for rather general fields), but only certain partial 2-trees guarantee maximum multiplicity 2. Among partial linear 2-trees, they are only those whose vertices can be covered b…
▽ More
Characterized are all simple undirected graphs $G$ such that any real symmetric matrix that has graph $G$ has no eigenvalues of multiplicity more than 2. All such graphs are partial 2-trees (and this follows from a result for rather general fields), but only certain partial 2-trees guarantee maximum multiplicity 2. Among partial linear 2-trees, they are only those whose vertices can be covered by two "parallel" induced paths. The remaining graphs that guarantee maximum multiplicity 2 are comprised by certain identified families of "exceptional" partial 2-trees that are not linear.
△ Less
Submitted 19 January, 2007;
originally announced January 2007.
-
Eigenvalues of Words in Two Positive Definite Letters
Authors:
Christopher J Hillar,
Charles R Johnson
Abstract:
The question of whether all words in two real positive definite letters have only positive eigenvalues is addressed and settled (negatively). This question was raised some time ago in connection with a long-standing problem in theoretical physics. A large class of words that do guarantee positive eigenvalues is identified, and considerable evidence is given for the conjecture that no other words…
▽ More
The question of whether all words in two real positive definite letters have only positive eigenvalues is addressed and settled (negatively). This question was raised some time ago in connection with a long-standing problem in theoretical physics. A large class of words that do guarantee positive eigenvalues is identified, and considerable evidence is given for the conjecture that no other words do. In the process, a fundamental question about solvability of symmetric word equations is encountered.
△ Less
Submitted 16 November, 2005;
originally announced November 2005.
-
Necessary And Sufficient Conditions For Existence of the LU Factorization of an Arbitrary Matrix
Authors:
Pavel Okunev,
Charles R. Johnson
Abstract:
If $A$ is an n-by-n matrix over a field $F$ ($A\in M_{n}(F)$), then $A$ is said to ``have an LU factorization'' if there exists a lower triangular matrix $L\in M_{n}(F)$ and an upper triangular matrix $U\in M_{n}(F)$ such that $$A=LU.$$ We give necessary and sufficient conditions for LU factorability of a matrix. Also simple algorithm for computing an LU factorization is given. It is an extensio…
▽ More
If $A$ is an n-by-n matrix over a field $F$ ($A\in M_{n}(F)$), then $A$ is said to ``have an LU factorization'' if there exists a lower triangular matrix $L\in M_{n}(F)$ and an upper triangular matrix $U\in M_{n}(F)$ such that $$A=LU.$$ We give necessary and sufficient conditions for LU factorability of a matrix. Also simple algorithm for computing an LU factorization is given. It is an extension of the Gaussian elimination algorithm to the case of not necessarily invertible matrices. We consider possibilities to factors a matrix that does not have an LU factorization as the product of an ``almost lower triangular'' matrix and an ``almost upper triangular'' matrix. There are many ways to formalize what almost means. We consider some of them and derive necessary and sufficient conditions. Also simple algorithms for computing of an ``almost LU factorization'' are given.
△ Less
Submitted 19 June, 2005;
originally announced June 2005.
-
Positive Eigenvalues of Generalized Words in Two Hermitian Positive Definite Matrices
Authors:
Christopher Hillar,
Charles R. Johnson
Abstract:
We define a word in two positive definite (complex Hermitian) matrices $A$ and $B$ as a finite product of real powers of $A$ and $B$. The question of which words have only positive eigenvalues is addressed. This question was raised some time ago in connection with a long-standing problem in theoretical physics, and it was previously approached by the authors for words in two real positive defini…
▽ More
We define a word in two positive definite (complex Hermitian) matrices $A$ and $B$ as a finite product of real powers of $A$ and $B$. The question of which words have only positive eigenvalues is addressed. This question was raised some time ago in connection with a long-standing problem in theoretical physics, and it was previously approached by the authors for words in two real positive definite matrices with positive integral exponents. A large class of words that do guarantee positive eigenvalues is identified, and considerable evidence is given for the conjecture that no other words do.
△ Less
Submitted 29 April, 2005;
originally announced April 2005.
-
Positive eigenvalues and two-letter generalized words
Authors:
Christopher Hillar,
Charles R. Johnson,
Ilya M. Spitkovsky
Abstract:
A generalized word in two letters $A$ and $B$ is an expression of the form $W=A^{α_1}B^{β_1}A^{α_2}B^{β_2}... A^{α_N}B^{β_N}$ in which the exponents $α_i$, $β_i$ are nonzero real numbers. When independent positive definite matrices are substituted for $A$ and $B$, we are interested in whether $W$ necessarily has positive eigenvalues. This is known to be the case when N=1 and has been studied in…
▽ More
A generalized word in two letters $A$ and $B$ is an expression of the form $W=A^{α_1}B^{β_1}A^{α_2}B^{β_2}... A^{α_N}B^{β_N}$ in which the exponents $α_i$, $β_i$ are nonzero real numbers. When independent positive definite matrices are substituted for $A$ and $B$, we are interested in whether $W$ necessarily has positive eigenvalues. This is known to be the case when N=1 and has been studied in case all exponents are positive by two of the authors. When the exponent signs are mixed, however, the situation is quite different (even for 2-by-2 matrices), and this is the focus of the present work.
△ Less
Submitted 28 April, 2005;
originally announced April 2005.
-
Absolutely Flat Idempotents
Authors:
Jonathan M. Groves,
Yonatan Harel,
Christopher J. Hillar,
Charles R. Johnson,
Patrick X. Rault
Abstract:
A real $n$-by-$n$ idempotent matrix $A$ with all entries having the same absolute value is called {\it absolutely flat}. We consider the possible ranks of such matrices and herein characterize the triples: size, constant, and rank for which such a matrix exists. Possible inequivalent examples of such matrices are also discussed.
A real $n$-by-$n$ idempotent matrix $A$ with all entries having the same absolute value is called {\it absolutely flat}. We consider the possible ranks of such matrices and herein characterize the triples: size, constant, and rank for which such a matrix exists. Possible inequivalent examples of such matrices are also discussed.
△ Less
Submitted 28 April, 2005;
originally announced April 2005.
-
Symmetric Word Equations in Two Positive Definite Letters
Authors:
Christopher J. Hillar,
Charles R. Johnson
Abstract:
A generalized word in two positive definite matrices A and B is a finite product of nonzero real powers of A and B. Symmetric words in positive definite A and B are positive definite, and so for fxed B, we can view a symmetric word, S(A,B), as a map from the set of positive definite matrices into itself. Given positive definite P, B, and a symmetric word, S(A,B), with positive powers of A, we de…
▽ More
A generalized word in two positive definite matrices A and B is a finite product of nonzero real powers of A and B. Symmetric words in positive definite A and B are positive definite, and so for fxed B, we can view a symmetric word, S(A,B), as a map from the set of positive definite matrices into itself. Given positive definite P, B, and a symmetric word, S(A,B), with positive powers of A, we defne a symmetric word equation as an equation of the form S(A,B) = P. Such an equation is solvable if there is always a positive definite solution A for any given B and P. We prove that all symmetric word equations are solvable. Applications of this fact, methods for solution, questions about unique solvability (injectivity), and generalizations are also discussed.
△ Less
Submitted 29 September, 2002;
originally announced September 2002.
