[Paper Review] Involutions in $S_n$ and associated coadjoint orbits

А. Н. Панов|ArXiv.org|Jan 19, 2008

Homotopy and Cohomology in Algebraic Topology3 references4 citations

TL;DR

This paper introduces a new family of coadjoint orbits in the Lie algebra of the unitriangular group $\mathrm{UT}(n,K)$ associated with involutions in the symmetric group $S_n$. By constructing explicit polarizations and deriving generators for the defining ideals of these orbits using minors of a characteristic matrix, the authors provide a complete description of orbit dimensions and ideal structures, extending known classifications beyond regular and subregular cases.

ABSTRACT

In the paper we study the coadjoint orbits of the group $\mathrm{UT}(n,K)$ associated with involutions. We obtain a formula for dimension of the orbit. We construct a polarization for the canonical element of orbit. We find a system of generators in the defining ideal of orbit.

Motivation & Objective

• To extend the classification of coadjoint orbits in the nilpotent Lie algebra $\mathfrak{ut}(n,K)$ beyond regular and subregular cases.
• To associate each involution $\sigma \in S_n$ with a family of coadjoint orbits $\Omega(f)$ for $f \in X_\sigma$.
• To construct a polarization $\mathfrak{p}_\sigma$ for any $f \in X_\sigma$, ensuring existence of primitive ideals in $U(\mathfrak{n})$.
• To determine the dimension of each orbit $\Omega(f)$ in terms of $l(\sigma) - s(\sigma)$, where $l(\sigma)$ is the number of simple reflections and $s(\sigma)$ the number of arbitrary reflections in a reduced decomposition of $\sigma$.
• To find explicit generators for the defining ideal $\mathcal{I}(\Omega(f))$, each of the form $P - c$ where $P$ is a minor coefficient of the characteristic matrix $\Phi(\tau)$.

Proposed method

• For each involution $\sigma \in S_n$, decompose it into commuting reflections $r_{\xi_1} \cdots r_{\xi_s}$ with $\xi_1 > \cdots > \xi_s$, forming the set $\mathcal{S} = \{\xi_1, \dots, \xi_s\}$.
• Define the set $\Pi_\sigma = \bigcup_{t=1}^{n-1} \Pi^{(t)}$, where $\Pi^{(t)} = \{\eta \in \Delta^{(t)} \mid \sigma_{t-1}(\eta) > 0\}$, and construct the polarization $\mathfrak{p}_\sigma$ as the span of $\{y_{it} \mid \varepsilon_t - \varepsilon_i \in \Pi_\sigma\}$.
• Introduce the formal matrix $\Phi$ and its characteristic matrix $\Phi(\tau) = \tau\Phi + E$, whose minors $M_I^J(\tau)$ are polynomials in $\tau$ with coefficients in $S(\mathfrak{n})$, and decompose them as $\tau^m(P_{I,0}^J + P_{I,1}^J\tau + \cdots)$.
• Use the Poisson bracket $\{y, P\}$ to derive recurrence relations between coefficients $P_{I,\mu}^J$ of minors, proving that $\{y, P_{I,\mu}^J\} \in \mathcal{I}(\Omega(f))$ via induction and admissible decompositions.
• Establish that the first non-zero coefficient $P_{I,0}^J$ of a minor corresponds to a minor of the matrix $\Phi$, and use this to generate elements $P - c$ in the ideal $\mathcal{I}(\Omega(f))$.
• Prove that the ideal $\mathcal{I}(\Omega(f))$ is generated by elements of the form $P - c$, where $P$ is a coefficient of a minor of $\Phi(\tau)$, and $c \in K$, using inductive arguments and properties of symmetric algebras.

Experimental results

Research questions

• RQ1 How can coadjoint orbits associated with involutions in $S_n$ be systematically constructed and classified in $\mathfrak{ut}(n,K)$?
• RQ2 What is the dimension of the coadjoint orbit $\Omega(f)$ for $f \in X_\sigma$, and how does it relate to the combinatorics of the involution $\sigma$?
• RQ3 Can a polarization $\mathfrak{p}_\sigma$ be explicitly constructed for each $f \in X_\sigma$, and is it maximal isotropic with respect to the skew form $f([x,y])$?
• RQ4 What is the structure of the defining ideal $\mathcal{I}(\Omega(f))$ of such orbits, and can it be generated by elements of the form $P - c$ where $P$ is a coefficient of a minor of $\Phi(\tau)$?
• RQ5 How do the coefficients of minors of $\Phi(\tau)$ transform under the Poisson bracket, and what does this imply for the ideal membership of $\{y, P\}$?

Key findings

• The polarization $\mathfrak{p}_\sigma$ constructed from the set $\Pi_\sigma$ is a maximal isotropic subspace for any $f \in X_\sigma$, thus providing a valid polarization for the orbit method.
• The dimension of the coadjoint orbit $\Omega(f)$ for $f \in X_\sigma$ is exactly $l(\sigma) - s(\sigma)$, where $l(\sigma)$ is the number of simple reflections and $s(\sigma)$ the number of arbitrary reflections in a reduced decomposition of $\sigma$.
• The defining ideal $\mathcal{I}(\Omega(f))$ is generated by elements of the form $P - c$, where $P$ is a coefficient of a minor of the characteristic matrix $\Phi(\tau)$, and $c \in K$, establishing a complete algebraic description of the orbit.
• The first non-zero coefficient $P_{I,0}^J$ of a minor $M_I^J(\tau)$ corresponds to a minor of the matrix $\Phi$, and this coefficient generates part of the ideal $\mathcal{I}(\Omega(f))$.
• The Poisson bracket $\{y, P\}$ of a coefficient $P$ of a minor lies in $\mathcal{I}(\Omega(f))$, which is proven via inductive decomposition of minors and comparison of coefficients in admissible expansions.
• The ideal $\mathcal{I}(\Omega(f))$ is generated by elements $P - c$ with $P$ being coefficients of minors of $\Phi(\tau)$, and this structure is preserved under the action of the Lie algebra, confirming the orbit's closedness and algebraic nature.