Leading coefficients of the Kazhdan-Lusztig polynomials for an Affine Weyl group of type $\widetilde{B_2}$

Liping WangarXiv 2008 · DOI · arXiv

Abstract

In this paper, we compute the leading coefficients $\mu(u,w)$ of the Kazhdan–Lusztig polynomials $P_{u,w}$ for an affine Weyl group of type $\widetilde{B}_2$. Utilizing the a-function of a Coxeter group as defined by Lusztig, we explicitly determine the majority of $\mu(u,w)$ values. These computations allow us to verify a long-standing conjecture by Lusztig regarding distinguished involutions for this specific group and identify a necessary modification to a previously proposed formula for algebraic coefficients.

Introduction

Since their introduction in the late 1970s, Kazhdan–Lusztig polynomials $P_{u,w}$ have become a cornerstone of modern representation theory. They provide deep insights into the internal structure of Lie algebras, quantum groups, and Hecke algebras. A parameter of particular importance is the leading coefficient, denoted as $\mu(u, w)$, which defines the edges of a W-graph. These graphs are combinatorial representations that describe how the Hecke algebra—a deformation of the group algebra—acts on its modules.

While $\mu$-values have been extensively studied for symmetric groups and affine Weyl groups of type $\widetilde{A}_n$, the affine Weyl group of type $\widetilde{B}_2$ (associated with specific rank-2 symmetries) has remained computationally elusive. This research fills that gap, providing the explicit values necessary to test broader conjectures regarding the geometric and algebraic properties of these groups.

Methodology

The computation of $\mu(u,w)$ for an infinite group like $\widetilde{B}_2$ requires a systematic partitioning of the group elements. The author employs several sophisticated algebraic tools:

  • Lusztig’s a-function: This function serves as a measure of "complexity" for group elements. By using it, the author partitions the group into four two-sided cells: $c_e, c_1, c_2,$ and $c_0$. This allows the computation to proceed cell-by-cell, reducing the global problem into manageable local segments.
  • Semilinear Equations: The research utilizes a system of equations introduced by Lusztig in 1996. These equations relate $\mu$-coefficients to the weight lattice of the root system, effectively transforming group-theoretic questions into a system of equations solvable via induction.
  • Star Operations: This standard technique in Coxeter group theory is used to transform elements while preserving specific polynomial properties, facilitating the mapping of known values to unknown ones.
  • Descending Induction: To address the most complex "lowest" cell ($c_0$), the author applies induction based on the partial ordering of group elements.

Key Insights and Theoretical Contributions

Validation of the Distinguished Involutions Conjecture

A primary achievement of this work is the proof of Theorem 2.2, which confirms that Lusztig’s 1987 conjecture on distinguished involutions holds for the affine Weyl group of type $\widetilde{B}_2$. This conjecture relates the set of elements $D_1$ (where the relationship between an element's length and its a-function value is specific) to the $\mu$-coefficients, reinforcing the structural consistency of Kazhdan–Lusztig theory across different group types.

Correction of the $b_{\lambda, \lambda'}$ Formula

The paper identifies a significant discrepancy in a formula proposed by Lusztig in 1996 (specifically formula (12) in [L3]) regarding the coefficients $b_{\lambda, \lambda'}$. Through explicit computation, the author provides a counterexample in type $\widetilde{B}_2$ and offers a corrected set of values in Proposition 5.1. This correction is vital for researchers utilizing Hecke algebra modules in higher-rank computations.

Non-Local Finiteness of the W-Graph

The study demonstrates that the W-graph for type $\widetilde{B}_2$ is non-locally finite. Unlike finite Weyl groups, a single element in this affine group can be connected to an infinite number of other elements through non-zero $\mu$-coefficients. This result highlights the inherent complexity of affine structures and the limitations of purely visual or manual graph construction.

Results

The paper provides exhaustive lists of $\mu(u, w)$ values across various cells:

  • Theorems 5.7, 5.8, and 5.9: Detail the $\mu$ values for interactions within and between specific cells.
  • Theorem 7.2 and 8.2: Characterize the leading coefficients for the interaction between the $c_2$ and $c_1$ cells, proving they are restricted to values of 0 or 1 under defined conditions.

Limitations and Future Directions

While the majority of $\mu$-values for $\widetilde{B}_2$ are now known, the "lowest" two-sided cell ($c_0$) remains only partially determined. Specifically, for cases where the a-function value equals 4, the author notes that only partial results are currently attainable.

Furthermore, the methods employed here are highly specific to the rank-2 case. Extending these explicit computations to higher-rank affine Weyl groups, such as $\widetilde{B}_n$ or $\widetilde{D}_n$, remains a significant theoretical and computational challenge for the field.

Summary updated

Abstract

In this paper we compute the leading coefficients $μ(u,w)$ of the Kazhdan--Lusztig polynomials $P_{u,w}$ for an affine Weyl group of type $\tilde{B}_2$. By using the \textbf{a}-function of a Coxeter group defined by Lusztig (see [L1, \S2]), we compute most $μ(u,w)$ explicitly. With part of these values $μ(u,w)$, we show that a conjecture of Lusztig on distinguished involutions is true for an affine Weyl group of type $\tilde{B}_2$. We also show that the conjectural formula in [L3, (12)] needs a modification.
BibTeX
@misc{wang2008leading,
  author = {Liping Wang},
  title = {{Leading coefficients of the Kazhdan-Lusztig polynomials for an Affine Weyl group of type $\textbackslash\{\}widetilde\{B\_2\}$}},
  year = {2008},
  eprint = {0805.3463},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/0805.3463},
}
Figure