Minimal dimension equivariant embeddings of real and complex flag manifolds into Euclidean spaces
Zhongzi Wang, and Hang Yin · 2026
Read Minimal dimension equivariant embeddings of real and complex flag manifolds into Euclidean spaces by Zhongzi Wang, and Hang Yin (2026) 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-2605-17337.txt.
Introduction
Wang and Yin determine the smallest equivariant embedding dimensions for orthogonal groups acting over real flag manifolds and unitary groups acting over complex ones. In both cases the minimum is attained at the isospectral model. The short paper settles the dimension question for these two families of group actions in one stroke.
Catalog record
| CREATOR | Zhongzi Wang, and Hang Yin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49050 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.17337
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Minimal dimension equivariant embeddings of real and complex flag manifolds into Euclidean spaces Zhongzi Wang, and Hang Yin Abstract . We determine the minimal equivariant embedding dimension of orthgonal groups acting on real flag manifolds and unitary groups acting on complex flag manifolds. The minimal embedding dimension is achieved at isospectral model. 1. Introduction A famous theorem in differential topology is any smooth n -dimensional manifold can be smoothly embedded into \mathbb{R}^{2n} [ 11 ] . There are many results concerning the embedding of manifolds see e.g. Hirsch [ 2 ] proved every orientable 3 -manifold smoothly embedded into \mathbb{R}^{5} , and actually the dimension 5 is the best possible for general 3-manifolds. The determination for the minimal dimension of smooth embedding of a specific manifold is in general a very hard problem, and there are many results on the estimate of embedding dimensions. We first recall the definition of equivariant embedding: Definition 1.1 . Suppose M is a smooth n -dimensional manifold and G be a Lie group acting smoothly on M . A G -equivariant embedding e is a smooth embedding e:M\to\mathbb{R}^{N} for some positive integer N together with a Lie group representation \rho:G\to{\rm O}(N) , such that for any g\in G , \rho(g)\circ e=e\circ g . Suppose M is a smooth compact manifold with a Lie group G -action. Mostow [ 4 ] , also Palais [ 5 ] , proved M has an G -equivariant embedding into \mathbb{R}^{N} for some positive integer N . We next recall the definition of a homogeneous manifold: Definition 1.2 . Suppose M is a smooth compact n -dimensional manifold and G be a compact Lie group acting smoothly on M . We say M is a homogeneous manifold if the G -action on M is transitive. The general problem is Question 1.3 . Suppose M is a smooth compact n -dimensional manifold and G be a compact Lie group acting smoothly on M . If M is a homogeneous manifold, let \text{ed}_{G}(M) be the minimal N such that there exists a G -equivariant embedding from M to \mathbb{R}^{N} . Determine the value of \text{ed}_{G}(M) . There are results concerning the upper bound for the dimension of equivariant embedding of manifolds see e.g. [ 10 ] . When G is a finite group, see also [ 9 ] . There has been studies on minimal dimension of equivariant embedding of finite group action on surfaces, see [ 7 ] , [ 6 ] . Recently, [ 3 ] determined the minimal dimension of equivariant embedding real flag manifolds with orthogonal group action except finitely many cases. In this paper, we determined the minimal dimension for equivariant embedding for arbitrary real flag manifolds and complex flag manifolds. The core method for studying \text{ed}_{G}(M) will be the use of the representation theory of Lie groups. Take a basepoint x_{0}\in M . Then the point stabilizer {\mathrm{Stab}}_{x_{0}}(G)=\{g\in G\mid g\cdot x_{0}=x_{0}\} is a closed subgroup of G , and we have an diffeomorphism M\cong G/{\mathrm{Stab}}_{x_{0}}(G) . The problem of equivariant embedding is equivalent to the problem of finding a representation space of G that contains a point with stabilizer equals to {\mathrm{Stab}}_{x_{0}}(G) . For representations of Lie groups, see [ 1 ] . For an n\times n -matrix T , denote T^{t} the trasnpose of T . If T\in M_{n}(\mathbb{C}) , denote T^{*} the conjugate transpose of T . 1.1. Isospectral model over \mathbb{R} Let \text{Sym}_{n}(\mathbb{R}) denote the space of symmetric n\times n -matrices over \mathbb{R} . Let \text{SO}_{n}(\mathbb{R}) denote the special orthogonal group. Note that \text{SO}_{n}(\mathbb{R}) acts on \text{Sym}_{n}(\mathbb{R}) by A\cdot M=AMA^{t} . So {\mathrm{Sym}}_{n}(\mathbb{R}) is a representation space of the special orthogonal group \text{SO}_{n}(\mathbb{R}) . The irreducible decomposition of this representation space is \text{Sym}_{n}(\mathbb{R})=\text{Sym}^{tr=0}_{n}(\mathbb{R})\oplus\mathbb{R} where \text{Sym}_{n}^{tr=0}(\mathbb{R})=\{T\in\text{Sym}_{n}(\mathbb{R}):trT=0\}. We say the representation space \text{Sym}_{n}(\mathbb{R}) is the isospectral model and the representation space \text{Sym}_{n}^{tr=0}(\mathbb{R}) is the traceless isospectral model. 1.2. Isospectral model over \mathbb{C} Let \text{Her}_{n}(\mathbb{R}) denote the space of Hermitian n\times n matrices, i. e. \text{Her}_{n}(\mathbb{R})=\{T\in M_{n}(\mathbb{C}):T=T^{*}\} . Let \text{SU}_{n} be the special unitary group. Note that \text{SU}_{n} acts on \text{Her}_{n}(\mathbb{R}) by A\cdot M=AMA^{*} , where A^{*}=\overline{A^{t}} . So \text{Her}_{n}(\mathbb{R}) is a representation space of the special unitary group \text{SU}_{n} . The irreducible decomposition of this representation space is \text{Her}_{n}(\mathbb{R})=\text{Her}^{tr=0}_{n}(\mathbb{R})\oplus\mathbb{R} where \text{Her}_{n}^{tr=0}(\mathbb{R})=\{T\in\text{Her}_{n}(\mathbb{R}):trT=0\}. We say the representation space \text{Her}_{n}(\mathbb{R}) is the isospectral model and the representation space \text{Her}_{n}^{tr=0}(\mathbb{R}) is the traceless isospectral model. 1.3. Real and complex flag manifolds The real flag manifold \mathrm{Flag}(n_{1},...,n_{r},\mathbb{C}^{n}) is a smooth manifold with a transitive \text{SO}_{n}(\mathbb{R}) -action. There is an {\mathrm{SO}}_{n}(\mathbb{R}) -equivariant embedding from \mathrm{Flag}(n_{1},...,n_{r}) into the traceless isospectral model {\mathrm{Sym}}_{n}^{tr=0}(\mathbb{R}). Let A be the block diagonal matrix of diagonal entries a_{1}I_{k_{1}},...,a_{r}I_{k_{r}} where a_{1},...,a_{r} are distinct real numbers with \sum_{i=1}^{r}a_{i}k_{i}=0 . The image of the {\mathrm{SO}}_{n}(\mathbb{R}) -equivariant embedding of the flag manifold \mathrm{Flag}(k_{1},...,k_{r};\mathbb{R}^{n}) is the {\mathrm{SO}}_{n} -orbit of A in {\mathrm{Sym}}_{n}^{tr=0}(\mathbb{R}). The complex flag manifold \mathrm{Flag}(n_{1},...,n_{r},\mathbb{C}^{n}) is a smooth complex manifold with a transitive {\mathrm{SU}}_{n} -action. There is an {\mathrm{SU}}_{n} -equivariant embedding from \mathrm{Flag}(n_{1},...,n_{r}) into the traceless isospectral model \text{Her}_{n}^{tr=0}(\mathbb{R}). Let A be the same matrix as in the real case. The image of the {\mathrm{SU}}_{n} -equivariant embedding of the flag manifold \mathrm{CFlag}(k_{1},...,k_{r};\mathbb{C}^{n}) is the SU_{n} -orbit of A in \text{Her}_{n}^{tr=0}(\mathbb{R}). The equivariant embedding for the real flag manifold is due to Lim-Ye. 1.4. Basic of flag variety Definition 1.4 . Let \mathbb{K} be a field. Let V be an n -dimensional vector space over \mathbb{K} . Let 0<k_{1}<...<k_{r}<n . (1) A flag of type (k_{1},...,k_{r}) in V is a filtration of linear subspaces W_{1}\subset...\subset W_{r}\subset V such that \text{dim}W_{i}=k_{i}. (2) When k_{i}=i , the flag is said to be a full flag . (3) Given 0<k_{1}<...<k_{r}<n , the set of all flags of type (k_{1},...,k_{r}) in V can be equipped with the structure of a smooth projective algebraic variety over \mathbb{K} . It is called the flag variety of type (k_{1},...,k_{r}) of V , and is denoted as \text{Flag}(k_{1},...,k_{r},V) . (4) When \mathbb{K}=\mathbb{R} or \mathbb{C} , we also call it a flag manifold of type (k_{1},...,k_{r}) . Now we assume \mathbb{K}=\mathbb{R} or \mathbb{C} . The flag manifold \mathrm{Flag}(k_{1},...,k_{r},V) has dimension \frac{n(n-1)}{2}-\sum_{i=1}^{r+1}\frac{l_{i}(l_{i}-1)}{2} where l_{1}=k_{1},l_{i}=k_{i}-k_{i-1} for 2\leq i\leq r and l_{r+1}=n-k_{r} . When r=1 and k_{1}=k , \mathrm{Flag}(k_{1},...,k_{r}) is the real Grassmann manifold Gr_{k}(\mathbb{R}^{n}) . The dimension of Gr_{k}(\mathbb{R}^{n}) is k(n-k) . It is known that for any inner product on V , {\mathrm{SO}}(V) acts transitively on \mathrm{Flag}(k_{1},\ldots,k_{r},V) , and the stabilizer in O(V) of a flag W_{1}\subset\cdots\subset W_{r} in \mathrm{Flag}(k_{1},\ldots,k_{r},V) is {\mathrm{O}}(U_{1})\times\cdots\times{\mathrm{O}}(U_{r+1}) . Here U_{i} is the orthogonal complement of W_{i-1} in W_{i} , and we identify W_{0}=0 , W_{r+1}=V . If we take stabilizer in {\mathrm{SO}}(V) , then we get {\mathrm{S}}({\mathrm{O}}(U_{1})\times\cdots\times{\mathrm{O}}(U_{r+1})) . Here the outer S means the total determinant is 1 . Therefore we have \text{Flag}(k_{1},...,k_{r})\cong\text{SO}_{n}(\mathbb{R})/H where H=\text{S}(\text{O}_{l_{1}}(\mathbb{R})\times\text{O}_{l_{2}}(\mathbb{R})\times...\times\text{O}_{l_{r+1}}(\mathbb{R})) . 1.5. Main theorem and preliminary discussions We are considering equivariant embedding of {\mathrm{SO}}_{n}({\mathbb{R}})/{\mathrm{S}}({\mathrm{O}}_{n_{1}}({\mathbb{R}})\times\cdots{\mathrm{O}}_{n_{r}}({\mathbb{R}})) , n_{1}+\cdots+n_{r}=n , n_{i}\geqslant 1 , n\geqslant 2 . This amounts to a finite dimensional real representation V of {\mathrm{SO}}_{n}({\mathbb{R}}) with a vector v\in V whose stabilizer is {\mathrm{S}}({\mathrm{O}}_{n_{1}}({\mathbb{R}})\times\cdots{\mathrm{O}}_{n_{r}}({\mathbb{R}})) . The isospectral model \{A\in{\mathrm{M}}_{n}({\mathbb{R}}):A\text{ symmetric and }{\mathrm{tr}}\ A=0\} is an equivariant embedding whose dimension is (n+2)(n-1)/2 . Our first main theorem is the following. Theorem 1.5 . The equivariant embedding of {\mathrm{SO}}_{n}({\mathbb{R}})/{\mathrm{S}}({\mathrm{O}}_{n_{1}}({\mathbb{R}})\times\cdots{\mathrm{O}}_{n_{r}}({\mathbb{R}})) , n_{1}+\cdots+n_{r}=n , n_{i}\geqslant