[Paper Review] Combinatorics of Character Formulas for the Lie Superalgebra $\fgl(m,n).$

Ian M. Musson, Vera Serganova|arXiv (Cornell University)|Apr 9, 2011

Algebraic structures and combinatorial models4 references4 citations

TL;DR

This paper provides a combinatorial proof of the equivalence between two distinct algorithms for computing composition factor multiplicities in Kac modules of the Lie superalgebra $gl(m,n)$, using weight and cap diagrams and path cancellation in a graph $ G$. The key contribution is establishing that non-split extensions between simple highest weight modules correspond exactly to edges of weight zero in a subgraph $ E$ of $ G$, resolving a long-standing problem in superalgebra representation theory.

ABSTRACT

Let $\fg$ be the Lie superalgebra $\fgl(m,n).$ Algorithms for computing the composition factors and multiplicities of Kac modules for $\fg$ were given by the second author in 1996, and by J. Brundan in 2003. We give a combinatorial proof of the equivalence between the two algorithms. The proof uses weight and cap diagrams introduced by Brundan and C. Stroppel, and cancelations between paths in a graph $\mathcal{G}$ defined using these diagrams. Each vertex of $\mathcal{G}$ corresponds to a highest weight of a finite dimensional simple module, and each edge is weighted by a nonnegative integer. If $\mathcal{E}$ is the subgraph of $\mathcal{G}$ obtained by deleting all edges of positive weight, then $\mathcal{E}$ is the graph that describes non-split extensions between simple highest weight modules. We also give a procedure for finding the composition factors of any Kac module, without cancelation. This procedure leads to a second proof of the main result.

Motivation & Objective

• To establish the equivalence between two combinatorial algorithms for computing composition factor multiplicities in Kac modules of $gl(m,n)$, as developed by Serganova and Brundan.
• To provide a new combinatorial proof of this equivalence using signed path sums in a graph $ G$ constructed from weight and cap diagrams.
• To characterize non-split extensions between finite-dimensional simple highest weight modules in terms of a subgraph $ E$ of $ G$, defined by removing edges of positive weight.
• To develop a cancellation-free procedure for computing composition factors of any Kac module, offering an alternative proof of the main result.
• To clarify the relationship between the representation theory of $gl(m,n)$ and Khovanov’s diagram algebra via categorical equivalence, as informed by prior work of Brundan and Stroppel.

Proposed method

• Define the free abelian group $ Z F$ with basis $F$, the set of functions from $ Z$ to $ { imes,igcirc,<,>}$ with finite support, encoding weight and cap diagrams.
• Construct a graph $ G$ where vertices represent highest weights of finite-dimensional simple modules, and edges are weighted by nonnegative integers based on diagram moves.
• Express Serganova’s formula as a signed sum over paths in $ G$, with each path corresponding to a term in $ Z F$.
• Define an involution on the set of paths that pairs terms of opposite sign, enabling cancellation to yield Brundan’s formula.
• Introduce the subgraph $ E riangleq G$ with all positive-weight edges removed, and show that $ Ext^1(L(f),L(g)) eq 0$ iff $f$ and $g$ are connected by an edge in $ E$.
• Use reverse induction on parabolic subalgebras $ q^{(i)}$ and the functor $ G^{(i)}$ to lift non-split extensions from lower to higher levels, proving the extension result.

Experimental results

Research questions

• RQ1 How can the two distinct combinatorial formulas for Kac module composition factor multiplicities—by Serganova and Brundan—be shown to be equivalent?
• RQ2 What is the precise combinatorial structure of the graph $ G$ that encodes the extension and multiplicity data of $gl(m,n)$ representations?
• RQ3 Which edges in $ G$ correspond to non-split extensions between simple highest weight modules?
• RQ4 Can a cancellation-free algorithm be constructed to compute composition factors of any Kac module?
• RQ5 How does the structure of weight and cap diagrams relate to Kazhdan-Lusztig polynomials and the representation theory of Khovanov’s diagram algebra?

Key findings

• The two algorithms for computing Kac module composition factor multiplicities—Serganova’s and Brundan’s—are combinatorially equivalent, as proven via path cancellation in the graph $ G$.
• The graph $ G$ encodes all composition factor multiplicities: each vertex corresponds to a highest weight of a finite-dimensional simple module, and edge weights represent multiplicity data.
• The subgraph $ E$, obtained by deleting all edges of positive weight from $ G$, classifies non-split extensions: $ Ext^1(L(f),L(g)) eq 0$ if and only if $f$ and $g$ are connected by an edge in $ E$.
• A new algorithm is constructed that computes composition factors without path cancellation, providing a second proof of the main equivalence result.
• The result is consistent with and can be derived from the equivalence of categories between the category $ F$ of finite-dimensional $ Z_2$-graded modules and modules over a version of Khovanov’s diagram algebra.
• A legal move of weight zero between diagrams corresponds precisely to a non-split extension, and such moves are characterized by interchanging labels at the endpoints of a cap in the cap diagram.