Simple polynomial equations over ( 2 × 2 ) (2\times 2) -matrices

Vitalij A. Chatyrko and Alexandre Karassev · 2025

Read Simple polynomial equations over ( 2 × 2 ) (2\times 2) -matrices by Vitalij A. Chatyrko and Alexandre Karassev (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-07689.txt.

Introduction

Vitalij Chatyrko and Alexandre Karassev solve polynomial equations whose unknown is a two-by-two matrix with real entries, in a math paper from 2025. The equation sets a matrix polynomial, built from powers of the unknown plus real coefficients and the identity matrix, equal to the zero matrix. Degrees run from two upward.

The solution set collects every matrix that satisfies a given equation. The authors describe that set completely, working in the natural Euclidean topology of the space of matrices. They also pin down its size in topological terms: the solution set always has dimension two. The paper thus gives a full picture, shape and dimension alike, for this basic class of matrix equations.

Catalog record

CREATORVitalij A. Chatyrko and Alexandre Karassev
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-07689

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Simple polynomial equations over ( 2 × 2 ) (2\times 2) -matrices Vitalij A. Chatyrko and Alexandre Karassev Abstract . We consider the polynomial equation X n + a n − 1 ⋅ X n − 1 + ⋯ + a 1 ⋅ X + a 0 ⋅ I = O , X^{n}+a_{n-1}\cdot X^{n-1}+\dots+a_{1}\cdot X+a_{0}\cdot I=O, over ( 2 × 2 ) (2\times 2) -matrices X X with the real entries, where I I is the identity matrix, O O is the null matrix, a i ∈ ℝ a_{i}\in\mathbb{R} for each i i and n ≥ 2 n\geq 2 . We discuss its solution set S S supplied with the natural Euclidean topology. We completely describe S S . We also show that dim S = 2 . \dim S=2. 2020 Mathematics Subject Classification Primary: 15A24, 54F45, Secondary: 15B30 1. Introduction As usual, by M_{2}(\mathbb{R}) we denote the algebra of real (2\times 2) -matrices, and by GL_{2}(\mathbb{R}) the group of invertible (2\times 2) -matrices. We equip M_{2}(\mathbb{R}) with any of equivalent norms that make M_{2}(\mathbb{R}) a topological algebra, homeomorphic to \mathbb{R}^{4}. Consider a polynomial with real coefficients f(x)=x^{n}+a_{n-1}x^{n-1}+\dots+a_{1}x+a_{0} with n\geq 1. It is well known that a polynomial equation [formula] f(x)=0 over the complex numbers \mathbb{C} has exactly n roots (counting multiplicities). Considered over the reals, the number of distinct roots of the equation (1) is either between 0 and n if n is even, or between 1 and n , if n is odd. Next, consider a polynomial F(X)=X^{n}+A_{n-1}X^{n-1}+\dots+A_{1}X+A_{0} of degree n\geq 1, where X and A_{i} , i=0,1,\dots,n-1, are from M_{2}(\mathbb{R}) , and a polynomial equation [formula] F(X)=O where O is the null (2\times 2) -matrix. Any matrix B\in M_{2}(\mathbb{R}) such that F(B)=O is called a solution of (2) . The solution set of (2) consists of all solutions of (2) . It is easy to check that the matrices \left(\begin{array}[]{rr}t&\sqrt{1-t^{2}}\\ \sqrt{1-t^{2}}&-t\end{array}\right),t\in[-1,1], are solutions of the equation X^{2}-I=O , where I is the identity (2\times 2) -matrix. Hence, for n=2 the number of distinct solutions of the equation (2) , unlike to the equation (1) over the real numbers, can be infinite (in this case, it has the cardinality of continuum). It is also readily seen that the equation X^{2}=\left(\begin{array}[]{rr}0&1\\ 0&0\end{array}\right) has no solution, the equation X^{2}=\left(\begin{array}[]{rr}1&1\\ 0&0\end{array}\right) has exactly 2 solutions, the equation X^{2}=\left(\begin{array}[]{rr}1&0\\ 0&2\end{array}\right) has exactly 4 solutions. Moreover, in [ W ] Wilson demonstrated a method showing that the polynomial equation (2) of degree 2 can also have exactly 1,3,5 or 6 solutions. The method uses eigenvalues and eigenvectors and it is due to Fuchs and Schwarz [ EFS ] . No other scenarios with finitely many solutions, except for those mentioned above, are possible. We simplify the polynomial F(X) by setting A_{m}=a_{m}\cdot I for each 0\leq m\leq n-1 , where each a_{m} is a real number. Denote the polynomial F(X) in this case by F_{s}(X) and note that F_{s}(X)=X^{n}+a_{n-1}\cdot X^{n-1}+\dots+a_{1}\cdot X+a_{0}\cdot I. Consider the equation [formula] F_{s}(X)=O In this note, we consider the following problem. Problem 1.1 . Describe the solution set S of (3) , where n\geq 2 . Everywhere in this note, by a space we mean a metric separable space. The covering dimension of a space X will be denoted by \dim X. See [ E , p. 385] for the definition of \dim and its main properties, such as: (i) The countable sum theorem [ E , Theorem 7.2.1] : if a space X is the union \cup_{i=1}^{\infty}X_{i} , where each X_{i} is a closed subset of X with \dim X_{i}\leq n,n\geq 0, then \dim X\leq n, and (ii) The monotone theorem [ E , Theorem 7.3.4] : for every subspace A of a space X we have \dim A\leq\dim X . Recall that \dim\mathbb{R}^{4}=4 and for any non-empty space X of cardinality less than continuum we have \dim X=0 . It is easy to see that the solution set S of (3) (and of (2) as well) is a closed subset of M_{2}(\mathbb{R}) and hence \dim S\leq 4 . In fact, we can easily improve this estimate as follows. Proposition 1.2 . \dim S\leq 3 . (The same is valid for the solution set of (2) .) Proof. Assume that some equation F_{s}(X)=0 , X\in M_{2}(\mathbb{R}), has the solution set S with \dim S=4 . Then there exists a non-empty open subset U of M_{2}(\mathbb{R}) such that U\subseteq S . Consider a matrix B(0)=\left(\begin{array}[]{rr}a&b\\ c&d\end{array}\right)\in U . Note that there exists \epsilon>0 such that the matrix B(t)=\left(\begin{array}[]{rr}(a+t)&b\\ c&d\end{array}\right) belongs to U for each 0\leq t\leq\epsilon . Further, F_{s}(B(t))=\left(\begin{array}[]{rr}t^{n}+p(t)&*\\ &*\end{array}\right) , where p(t) is a polynomial of degree \leq n-1 , and t^{n}+p(t)=0 for every 0\leq t\leq\epsilon . But the last is impossible. We have a contradiction. Hence, \dim S\leq 3 . ∎ In connection with this, one can ask the following. Question 1.3 . What is \dim S ? In this note we will solve Problem 1.1 (see Theorem 3.4 ) and answer Question 1.3 (see Theorem 4.5 ). 2. Auxiliary facts Recall that (2\times 2) -matrices A and B with entries from a field \mathbb{F} are similar over \mathbb{F} if there exists a (2\times 2) -invertible matrix C with entries from \mathbb{F} such that A=C^{-1}BC , and the similarity is an equivalence relation. For real numbers p and q , let D(p,q)=\left(\begin{array}[]{rr}p&0\\ 0&q\end{array}\right),J(p)=\left(\begin{array}[]{rr}p&1\\ 0&p\end{array}\right), and for real numbers a and b with a^{2}+b^{2}\neq 0 , let R(a,b)=\left(\begin{array}[]{rr}a&b\\ -b&a\end{array}\right),U(a,b)=\left(\begin{array}[]{rr}a+ib&0\\ 0&a-ib\end{array}\right). The following propositions are standard facts from matrix algebra (see, for example, [ L ] ). Proposition 2.1 . For any (2\times 2) -matrix A\in M_{2}(\mathbb{R}) there exists an invertible C\in M_{2}(\mathbb{R}) such that one of the following is true: A=C^{-1}D(p,q)C , A=C^{-1}J(p)C, or A=C^{-1}R(a,b)C, i.e. A is similar over \mathbb{R} either D(p,q) , J(p) or R(a,b). Proposition 2.2 . Any A\in M_{2}(\mathbb{R}) is simiilar to R(a,b) over \mathbb{R} iff A is similar to U(a,b) over \mathbb{C} . Let f_{s}(x)=x^{n}+a_{n-1}x^{n-1}+\dots+a_{1}x+a_{0} be a polynomial with real coefficients a_{i},0\leq i\leq n-1, corresponding to the polynomial F_{s}(X) . Consider a polynomial equation [formula] f_{s}(x)=0. The following lemmas are evident. Lemma 2.3 . * F_{s}(B(p,q))=\left(\begin{array}[]{rr}f_{s}(p)&0\\ 0&f_{s}(q)\end{array}\right) ; F_{s}(J(p))=\left(\begin{array}[]{rr}f_{s}(p)&f^{\prime}_{s}(p)\\ 0&f_{s}(p)\end{array}\right) ; F_{s}(U(a,b))=\left(\begin{array}[]{rr}f_{s}(a+ib)&0\\ 0&f_{s}(a-ib)\end{array}\right) . In particular, p and q are roots of (4) iff B(p,q) is a solution of (3) ; p is a root of (4) with the multiplicity \geq 2 iff J(p) is a solution of (3) ; a\pm ib are roots of (4) iff F_{s}(U(a,b))=O. Lemma 2.4 . If (2\times 2) -matrices A and B (with real or complex entries) are similar over \mathbb{R} or \mathbb{C} , and F_{s}(A)=O, then F_{s}(B)=O. In particular, R(a,b) is a solution of (3) iff a\pm ib are roots of (4) . 3. The structure of S For a given B\in M_{2}(\mathbb{R}) , define a map Conj_{B}\colon GL_{2}(\mathbb{R})\to M_{2}(\mathbb{R}) by Conj_{B}(X)=X^{-1}BX . The following is easy to verify. Lemma 3.1 . The map Conj_{B} is continuous, and the image Conj_{B}(GL_{2}(\mathbb{R})) is \sigma -compact. For real numbers p , q , and real nubers a , b with a^{2}+b^{2}\neq 0 , let S(D(p,q))=Conj_{D(p,q)}(GL_{2}(\mathbb{R})), S(J(p))=Conj_{J(p)}(GL_{2}(\mathbb{R})), S(R(a,b))=Conj_{R(a,b)}(GL_{2}(\mathbb{R})). Remark 3.2 . (i) For any p\in\mathbb{R} we have S(D(p,p))=\{D(p,p)\} and \dim S(D(p,p))=0 . (ii) For any real numbers p\neq q and a matrix C(\phi)=\left(\begin{array}[]{rr}\cos\phi&\sin\phi\\ -\sin\phi&\cos\phi\end{array}\right) we have C(\phi)^{-1}B(p,q)C(\phi)=\left(\begin{array}[]{rr}a(p,q,\phi)&b(p,q,\phi)\\ c(p,q,\phi)&d(p,q,\phi)\end{array}\right), where a(p,q,\phi)=p\cos^{2}(\phi)+q\sin^{2}\phi

Read the complete text