Abstract
A Fan–Theobald–von Neumann system (Gowda in Optimizing certain combinations of linear/distance functions over spectral sets, 2019. arXiv:1902.06640v2) is a triple \(({{\mathcal {V}}},{{\mathcal {W}}},\lambda )\), where \({{\mathcal {V}}}\) and \({{\mathcal {W}}}\) are real inner product spaces and \(\lambda :{{\mathcal {V}}}\rightarrow {{\mathcal {W}}}\) is a norm-preserving map satisfying a Fan–Theobald–von Neumann type inequality together with a condition for equality. Examples include Euclidean Jordan algebras, systems induced by certain hyperbolic polynomials, and normal decompositions systems (Eaton triples). In Gowda (Optimizing certain combinations of linear/distance functions over spectral sets, 2019. arXiv:1902.06640v2), we presented some basic properties of such systems and described results on optimization problems dealing with certain combinations of linear/distance and spectral functions. We also introduced the concept of commutativity via the equality in the Fan–Theobald–von Neumann-type inequality. In the present paper, we elaborate on the concept of commutativity and introduce/study automorphisms, majorization, and reduction in Fan–Theobald–von Neumann systems.
Similar content being viewed by others
Data availability
Data sharing not applicable to this article as no data sets were generated or analyzed during the current study.
References
Alberti, P.M., Uhlmann, A.: Stochasticity and Partial Order. Dt. Verlag Wissenshaften, Berlin (1982)
Ando, T.: Majorization, doubly stochastic matrices, and comparison of eigenvalues. Linear Algebra Appl. 118, 163–248 (1989)
Baes, M.: Convexity and differentiability properties of spectral functions in Euclidean Jordan algebras. Linear Algebra Appl. 422, 664–700 (2007)
Bauschke, H.H., Güler, O., Lewis, A.S., Sendov, H.S.: Hyperbolic polynomials and convex analysis. Can. J. Math. 53, 470–488 (2001)
Bennett, C., Sharpley, R.C.: Interpolation of Operators. Academic Press, Cambridge (1988)
Bhatia, R.: Matrix Analysis. Springer, New York (1997)
Eaton, M.L.: On group induced orderings, monotone functions, and convolution theorems. In: Tong, Y.L. (ed.) Inequalities in Statistics and Probability. IMS Lectures Notes, Monograph Series 5, pp. 13–25. Hayward, IMS (1984)
Eaton, M.L.: Group induced orderings with some applications in statistics. CWI Newsl. 16, 3–31 (1987)
Eaton, M.L., Perlman, M.D.: Reflection groups, generalized Schur functions, and the geometry of majorization. Ann. Probab. 5, 829–860 (1977)
Fan, K.: On a theorem of Weyl concerning eigenvalues of linear transformations I. Proc. Natl. Acad. Sci. USA 35, 652–655 (1949)
Faraut, J., Korányi, A.: Analysis on Symmetric Cones. Clarendon Press, Oxford (1994)
Gowda, M.S.: Optimizing certain combinations of linear/distance functions over spectral sets. arXiv:1902.06640v2 (2019)
Gowda, M.S.: Positive and doubly stochastic maps, and majorization in Euclidean Jordan algebras. Linear Algebra Appl. 528, 40–61 (2017)
Gowda, M.S.: Commutation principles for optimization problems on spectral sets in Euclidean Jordan algebras. Opt. Lett. 16, 1119–1128 (2022). arXiv:2009.04874v1
Gowda, M.S., Jeong, J.: Commutation principles in Euclidean Jordan algebras and normal decomposition principles. SIAM J. Optim. 27, 1390–1402 (2017)
Gowda, M.S., Tao, J.: The Cauchy interlacing theorem in simple Euclidean Jordan algebras and some consequences. Linear Multilinear Algebra 59, 65–86 (2011)
Gurvits, L.: Combinatorics hidden in hyperbolic polynomials and related topics. arXiv:math/0402088v1 (2004)
Jeong, J., Gowda, M.S.: Spectral cones in Euclidean Jordan algebras. Linear Algebra Appl. 509, 286–305 (2016)
Jeong, J., Gowda, M.S.: Spectral sets and functions in Euclidean Jordan algebras. Linear Algebra Appl. 518, 31–56 (2017)
Lewis, A.S.: Group invariance and convex matrix analysis. SIAM J. Matrix Anal. 17, 927–949 (1996)
Lewis, A.S.: Convex analysis on Cartan subspaces, Nonlinear. Analysis 42, 813–820 (2000)
Lewis, A.S., Parrilo, P.A., Ramana, M.V.: The Lax conjecture is true. Proc. Am. Math. Soc. 133, 2495–2499 (2005)
Lim, Y., Kim, J., Faybusovich, L.: Simultaneous diagonalization on simple Euclidean Jordan algebras and its applications. Forum Math. 15, 639–644 (2003)
Marshall, A.W., Olkin, I.: Inequalities: Theory of Majorization and its Applications. Academic Press, New York (1979)
Mirsky, L.: Trace of matrix products. Math. Nach. 20, 171–174 (1959)
Niezgoda, M.: Group majorization and Schur type inequalities. Linear Algebra Appl. 268, 9–30 (1998)
Niezgoda, M.: Extended commutation principles for normal decomposition systems. Linear Algebra Appl. 539, 251–273 (2018)
Orlitzky, M.: The NDS-FTvN connection, private communication (June 20, 2019)
Orlitzky, M.: Proscribed normal decomposition systems of Euclidean Jordan algebras. J. Convex Anal. 29, 755–766 (2022)
Richter, H.: Zur abschätsung von matrizennormen. Math. Nach. 18, 178–187 (1958)
Rockafellar, R.T.: Convex Analysis. Princeton University Press, Princeton (1970)
Rudin, W.: Functional Analysis. McGraw-Hill, New York (1973)
Tam, T.-Y.: An extension of a result of Lewis. Electron. J. Linear Algebra 5, 1–10 (1999)
Theobald, C.M.: An inequality for the trace of the product of two symmetric matrices. Math. Proc. Camb. Philos. Soc. 77, 265–267 (1975)
von Neumann, J.: Some matrix inequalities and metrization of matric-space, Tomsk University Rev., 1 (1937) 286–300; In Collected Works, Vol. IV, Pergamon, Oxford, 205–218 (1962)
Acknowledgements
We wish to record our appreciation to Roman Sznajder and Michael Orlitzky for their comments and suggestions on an earlier draft of the paper. The work of J. Jeong was supported by the National Research Foundation of Korea NRF-2016R1A5A1008055 and the National Research Foundation of Korea NRF-2021R1C1C2008350.
Funding
Juyoung Jeong was supported by the National Research Foundation of Korea NRF-2016R1A5A1008055 and the National Research Foundation of Korea NRF-2021R1C1C2008350.
Author information
Authors and Affiliations
Contributions
The authors equally contributed to formulation, literature search, and writhing and reviewing the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors have no relevant financial or non-financial interests to disclose.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
1.1 Normal Decomposition Systems
Definition 10.6
[20]. Let \({{\mathcal {V}}}\) be a real inner product space, \({{\mathcal {G}}}\) be a closed subgroup of the orthogonal group of \({{\mathcal {V}}}\), and \(\gamma : {{\mathcal {V}}}\rightarrow {{\mathcal {V}}}\) be a map satisfying the following conditions:
-
(a) \(\gamma \) is \({{\mathcal {G}}}\)-invariant, that is, \(\gamma (Ax) = \gamma (x)\) for all \(x \in {{\mathcal {V}}}\) and \(A \in {{\mathcal {G}}}\).
-
(b) For each \(x\in {{\mathcal {V}}}\), there exists \(A\in {{\mathcal {G}}}\) such that \(x=A\gamma (x)\).
-
(c) For all \(x,y\in {{\mathcal {V}}}\), we have \(\langle x, y \rangle \le \langle \gamma (x), \gamma (y) \rangle \).
Then, \(({{\mathcal {V}}}, {{\mathcal {G}}}, \gamma )\) is called a normal decomposition system.
Items (a) and (b) in the above definition show that \(\gamma ^2=\gamma \) and \(\left\| \gamma (x) \right\| = \left\| x \right\| \) for all x. We state a few relevant properties.
Proposition 10.7
[20, Proposition 2.3 and Theorem 2.4]. Let \(({{\mathcal {V}}},{{\mathcal {G}}},\gamma )\) be a normal decomposition system. Then,
-
(i)
For any two elements x and y in \({{\mathcal {V}}}\), we have
$$\begin{aligned} \max _{A \in {{\mathcal {G}}}}\, \langle Ax, y \rangle = \langle \gamma (x), \gamma (y) \rangle . \end{aligned}$$Also, \(\langle x, y \rangle = \langle \gamma (x), \gamma (y) \rangle \) holds for two elements x and y if and only if there exists an \(A \in {{\mathcal {G}}}\) such that \(x = A\gamma (x)\) and \(y = A\gamma (y)\).
-
(ii)
The range of \(\gamma \), denoted by F, is a closed convex cone in \({{\mathcal {V}}}\).
1.2 Eaton Triples
Eaton triples were introduced and studied in [7,8,9] from the perspective of majorization techniques in probability. They were also extensively studied in the papers of Tam and Niezgoda, see the references.
Definition 10.8
Let \({{\mathcal {V}}}\) be a finite dimensional real inner product space, \({{\mathcal {G}}}\) be a closed subgroup of the orthogonal group of \({{\mathcal {V}}}\), and F be a closed convex cone in \({{\mathcal {V}}}\) satisfying the following conditions:
-
(a) \(\textrm{Orb}(x)\cap F \ne \emptyset \) for all \(x \in {{\mathcal {V}}}\), where \(\textrm{Orb}(x) := \{Ax : A \in {{\mathcal {G}}}\}\).
-
(b) \(\langle x, Ay \rangle \le \langle x, y \rangle \) for all \(x, y \in F\) and \(A \in {{\mathcal {G}}}\).
Then, \(({{\mathcal {V}}}, {{\mathcal {G}}}, F)\) is called an Eaton triple.
It has been shown (see [26], page 14) that in an Eaton triple \((V, {{\mathcal {G}}}, F)\), \(\text {Orb}(x)\cap F\) consists of exactly one element for each \(x\in {{\mathcal {V}}}\). Defining \(\gamma : {{\mathcal {V}}}\rightarrow {{\mathcal {V}}}\) such that \(\text {Orb}(x) \cap F = \{\gamma (x)\}\), it has been observed that \(({{\mathcal {V}}}, {{\mathcal {G}}}, \gamma )\) is a normal decomposition system. Also, given a finite dimensional normal decomposition system \(({{\mathcal {V}}},{{\mathcal {G}}}, \gamma )\) with \(F := \gamma ({{\mathcal {V}}})\), \(({{\mathcal {V}}}, {{\mathcal {G}}}, F)\) becomes an Eaton triple. Thus, finite dimensional normal decomposition systems are equivalent to Eaton triples [20, 21, 27].
1.3 Rearrangement Inequality for Measurable Functions
The notion of rearrangement of a function, systematically introduced by Hardy and Littlewood, has played a key role in proving inequalities in classical and applied analysis. The definitions and properties in this subsection can be found in [5], Chapter 2.
Let \((\Omega , \Sigma , \mu )\) denote a \(\sigma \)-finite measure space.
Definition 10.9
Let \(f : \Omega \rightarrow {\mathcal {R}}\) be a \(\Sigma \)-measurable function.
-
The function \(\mu _f : [0, \infty ) \rightarrow [0, \infty ]\) defined by
$$\begin{aligned} \mu _f(\alpha ) = \mu \big ( \{x \in \Omega : \left| f(x) \right| > \alpha \} \big ) \end{aligned}$$is called the distribution function of f.
-
The decreasing rearrangement of f is the function \(f^* : [0, \infty ) \rightarrow [0, \infty ]\) defined by
$$\begin{aligned} f^*(t) := \inf \{ \alpha \ge 0 : \mu _f(\alpha ) \le t \}, \end{aligned}$$where we use the convention that \(\inf \emptyset = \infty \).
In the next two propositions, we record some basic properties of the distribution function and decreasing rearrangement.
Proposition 10.10
The following properties hold:
-
(i)
f and \(f^*\) are equimeasurable, that is,
$$\begin{aligned} \mu \big ( \{x \in \Omega : \left| f(x) \right|> \alpha \} \big ) = m \big ( \{t> 0 : f^*(t) > \alpha \} \big ) \end{aligned}$$for all \(\alpha \ge 0\), where m is the Lebesgue measure.
-
(ii)
\(\left\| f \right\| _{L_{p}(\Omega )} = \left\| f^* \right\| _{L_{p}[0, \infty )}\) for all positive real numbers p.
-
(iii)
The Hardy–Littlewood–Pólya inequality holds, i.e.,
$$\begin{aligned} \int _{\Omega } \left| fg \right| \; d\mu \le \int _{0}^{\infty } f^{*}(t)g^{*}(t) \; dt. \end{aligned}$$
Proposition 10.11
There exists only one right-continuous decreasing function \(f^*\) equimeasurable with f. Hence, the decreasing rearrangement is unique.
Example 10.12
Consider the measure space \((\Omega , \Sigma , \mu ) = ({\mathbb {N}}, 2^{{\mathbb {N}}}, \mu )\), where \(\mu \) is the counting measure on \({\mathbb {N}}\). Then, any \(\Sigma \)-measurable function \(f : \Omega \rightarrow {\mathcal {R}}\) can be realized as a sequence \((x_n)\). The decreasing rearrangement \(f^*\) is a function defined on \([0, \infty )\), but can be interpreted as a sequence \((x_n^*)\), where, for any \(n\in {\mathbb {N}}\),
Formally, for any n,
The following can easily be observed:
-
(a) If \((x_n)\) has finite number of nonzero entries, say, k of them, then \(x^*=(\left| x_{n_1} \right| , \left| x_{n_2} \right| , \ldots , \left| x_{n_k} \right| , 0, 0, \ldots )\) with \(\left| x_{n_1} \right| \ge \left| x_{n_2} \right| \ge \cdots \ge \left| x_{n_k} \right| \).
-
(b) If \((x_n)\) has infinitely many nonzero entries, then \(x^*\) consists of absolute values of these entries arranged in the decreasing order; in particular, every entry of \(x^*\) is nonzero.
For example, if \(x = \Big ( 1, 0, \frac{1}{2}, 0, \frac{1}{3}, 0, 0, \ldots \Big )\), then \(x^* = \Big ( 1, \frac{1}{2}, \frac{1}{3}, 0, 0, \ldots \Big )\). On the other hand, if \(x = \Big ( 1, 0, \frac{1}{2}, 0, \frac{1}{3}, 0, \frac{1}{4}, 0, \ldots \Big )\), then \(x^* = \Big ( 1, \frac{1}{2}, \frac{1}{3}, \frac{1}{4}, \ldots \Big )\).
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Gowda, M.S., Jeong, J. Commutativity, Majorization, and Reduction in Fan–Theobald–von Neumann Systems. Results Math 78, 72 (2023). https://doi.org/10.1007/s00025-023-01845-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00025-023-01845-2
Keywords
- Fan–Theobald–von Neumann system
- Euclidean Jordan algebra
- normal decomposition system
- Eaton triple
- hyperbolic polynomial
- spectral set
- eigenvalue map
- strong operator commutativity