Real linear operators and numerical ranges
Damian Kołaczek, Vladimir Müller · 2025
Read Real linear operators and numerical ranges by Damian Kołaczek, Vladimir Müller (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-2508-04347.txt.
Introduction
Damian Kołaczek and Vladimir Müller survey a broad class of operators in this 2025 math paper. Real linear operators act between complex spaces and cover two families at once: plain linear maps and their mirror-image cousins. The paper gathers what is known about their shape, spectrum, and pairing properties.
The main new topic is the numerical range, the set of values an operator produces when fed unit inputs. For the classical case this set is known to be convex, shaped like a filled oval. The authors show the range stays convex for real linear operators on Hilbert spaces with at least two dimensions. The result extends a century-old theorem to the wider family.
Catalog record
| CREATOR | Damian Kołaczek, Vladimir Müller |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 65 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-04347
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Real linear operators and numerical ranges Thanks: The second author was supported by grant no. 25-15444K of GA ČR and RVO:67985840 Damian Kołaczek, Vladimir Müller Abstract . Real linear operators between two complex Banach spaces unify naturally two important classes of linear operators and antilinear operators. We give a survey of basic geometric, spectral and duality properties of real linear operators. The main goal of the paper is to introduce the numerical range of real linear operators on both Hilbert and Banach spaces and to study its properties. In particular, we show that the numerical range of real linear operators on complex Hilbert space is always a convex set (on at least two-dimensional spaces). This generalizes the classical result of Hausdorff and Toeplitz for linear operators. 1991 Mathematics Subject Classification Primary 47A12; Secondary 47A05 1. Introduction Let X,Y be complex Banach spaces. A continuous mapping \Phi:X\to Y is called a real linear operator if \Phi(x+x^{\prime})=\Phi(x)+\Phi(x^{\prime}) and \Phi(tx)=t\Phi(x) for all x,x^{\prime}\in X and t\in\mathbb{R} . Real linear operators appear frequently in various applications, see e.g. [ 14 ] , [ 13 ] , [ 22 ] , [ 2 ] , [ 28 ] and the references therein. A survey of the properties of real linear operators on complex Hilbert spaces was given in [ 20 ] . The most important examples of real linear operators are linear operators and antilinear operators. A continuous mapping \Phi:X\to Y is called an antilinear operator if \Phi(x+x^{\prime})=\Phi(x)+\Phi(x^{\prime}) and \Phi(\lambda x)=\bar{\lambda}\Phi(x) for all x,x^{\prime}\in X and \lambda\in\mathbb{C} . Some properties of antilinear operators are analogous to the properties of linear operators, but there are also some significant differences. For various aspects of antilinear operators, see e.g. [ 32 ] , [ 18 ] , [ 7 ] , [ 8 ] , [ 24 ] , [ 25 ] , [ 29 ] . Each real linear operator can be written in a canonical way as a sum of a linear operator and an antilinear operator. In the present paper, we continue the study of real linear operators. The paper is organized as follows. In the second section, we summarize some basic facts concerning real linear operators. In the next section, we show that the linear operators and antilinear operators are Birkhoff-James orthogonal in the Banach space of all real linear operators (between two given Banach spaces). The set of all real linear operators on a Banach space is a real Banach algebra. The spectrum of real linear operators on complex Hilbert spaces was introduced and studied in [ 20 ] and [ 31 ] . The most important difference from the spectral theory of linear operators is that the spectrum may be empty. In the fourth section of the present paper, we summarize the basic properties of spectrum of real linear operators on Banach spaces. The main goal of the paper is to introduce and study the notion of numerical range for real linear operators in the general case. For the special case of real linear operators on Hilbert space \mathbb{C}^{n} the notion of numerical range was already introduced in [ 17 ] . The numerical range of linear operators is already a classical concept and plays an important role in operator theory. The standard reference are two manuscripts of Bonsall and Duncan [ 6 ] , [ 5 ] . For a more recent survey, see [ 15 ] . The numerical range was recently extended and studied in the antilinear setting by a number of papers, see [ 10 ] , [ 16 ] , [ 26 ] . In the present paper, we introduce and study the numerical range of real linear operators both on Hilbert and on Banach spaces. We prove the basic properties of the numerical range and its relation with the spectrum and duality. Not surprisingly, the numerical range has nicer properties in the Hilbert space setting. We show that if the complex Hilbert space is at least two-dimensional then the numerical range of a real linear operator is always convex subset of the complex plane. This generalizes the famous Hausdorff-Toeplitz theorem for linear operators. It is well known that the numerical radius w(T) of a linear operator T satisfies the inequality w(T)\geq\|T\|/2. If T is self-adjoint, then w(T)=\|T\| . For antilinear and real linear operators, the situation is more complicated. We study to what extent such results can be generalized to the real linear setting. Another important result in the theory of the numerical range of linear operators is that the numerical range of a 2\times 2 -matrix is always an elliptical disc (possibly degenerated to a line segment or a single point). We show that this is no longer true for real linear operators. The numerical range of a real linear operator on a 2 -dimensional complex Hilbert space is always convex, but it can be a more complicated set. In the last section, we give some examples of this kind obtained numerically by a computer. 2. Basic results Let X,Y be complex Banach spaces. A mapping \Phi:X\to Y is called real linear operator if it is continuous, \Phi(x+x^{\prime})=\Phi(x)+\Phi(x^{\prime}) and \Phi(tx)=t\Phi(x) for all x,x^{\prime}\in X and real t . It is well known that a real linear mapping is continuous if and only if it is bounded. For a real linear operator \Phi:X\to Y we will write \|\Phi\|=\sup\{\|\Phi(x)\|:x\in X,\|x\|\leq 1\} . Denote by {\mathcal{R}}(X,Y) the set of all continuous real linear operators from X to Y . If Y=X , then we write shortly {\mathcal{R}}(X)={\mathcal{R}}(X,X) . In fact, a continuous additive mapping is automatically real linear. This observation is already classical, see [ 4 ] . Theorem 2.1 . Let \Phi:X\to Y be a continuous additive mapping (i.e., \Phi(x+x^{\prime})=\Phi(x)+\Phi(x^{\prime}) for all x,x^{\prime}\in X ). Then \Phi(tx)=t\Phi(x) for all x\in X and real t . Proof. It is easy to see that \Phi(nx)=n\Phi(x) for all x\in X and n\in\mathbb{N} . Consequently, \Phi(tx)=t\Phi(x) for all x\in X and rational t . Using the continuity of \Phi , we get \Phi(tx)=t\Phi(x) for all x\in X and the real number t . ∎ Moreover, any bounded additive mapping is continuous at the origin, and so continuous, see [ 4 ] . Basic examples of real linear operators are linear operators and antilinear operators. In fact, each real linear operator can be written as a sum of a linear operator and an antilinear operator, cf. [ 20 ] . Theorem 2.2 . Let \Phi:X\to Y be a real linear operator. Then there exist a linear operator T:X\to Y and an antilinear operator A:X\to Y such that \Phi=T+A . The operators T and A are determined uniquely. Proof. Define T and A by Tx=\frac{1}{2}\bigl(\Phi(x)-i\Phi(ix)\bigr) and Ax=\frac{1}{2}\bigl(\Phi(x)+i\Phi(ix)\bigr) for all x\in X . It is easy to see that \Phi=T+A , T is a linear operator and A an antilinear operator. If \Phi=T^{\prime}+A^{\prime} for another pair T^{\prime},A^{\prime} with T^{\prime} linear and A^{\prime} antilinear, then (T-T^{\prime})+(A-A^{\prime})=0 . So for each x\in X we have (T-T^{\prime})x+(A-A^{\prime})x=0 and 0=(T-T^{\prime})(ix)+(A-A^{\prime})(ix)=i(T-T^{\prime})x-i(A-A^{\prime})x, and so (T-T^{\prime})x-(A-A^{\prime})x=0 . Hence (T-T^{\prime})x=0 and (A-A^{\prime})x=0 . Since x\in X was arbitrary, we have T^{\prime}=T and A^{\prime}=A . ∎ Denote by X^{*} and Y^{*} the dual spaces of X and Y , respectively. For a linear