[Paper Review] An inhomogeneous Lambda-determinant

Philippe Di Francesco|arXiv (Cornell University)|Sep 28, 2012

Algebraic structures and combinatorial models13 references4 citations

TL;DR

This paper introduces a multi-parameter, inhomogeneous generalization of the Lambda-determinant using a deformed Dodgson condensation algorithm with two sets of coefficients $\lambda_a$ and $\mu_a$. It proves a closed-form formula expressing the generalized determinant as a weighted sum over alternating sign matrices (ASMs), where weights depend on matrix entries and vertex types in the 6V model, extending the classical Robbins-Rumsey result to inhomogeneous parameters and linking it to cluster algebras and network models.

ABSTRACT

We introduce a multi-parameter generalization of the Lambda-determinant of Robbins and Rumsey, based on the cluster algebra with coefficients attached to a T-system recurrence. We express the result as a weighted sum over alternating sign matrices.

Motivation & Objective

• To generalize the classical Lambda-determinant to an inhomogeneous, multi-parameter version using coefficients $\lambda_a$ and $\mu_a$.
• To establish the generalized determinant as a solution of a $T$-system with inhomogeneous coefficients, rooted in cluster algebra theory.
• To derive a closed-form expression for the generalized determinant as a sum over $n \times n$ alternating sign matrices (ASMs), with explicit weights.
• To connect the generalized determinant to statistical mechanics via the 6V model with domain wall boundary conditions.
• To explore the limit shape and singularity structure of the generating function in the inhomogeneous case, extending known results on Arctic circles.

Proposed method

• Define the generalized Lambda-determinant inductively via a modified Dodgson condensation algorithm using shift operators on $\lambda_a$ and $\mu_a$.
• Use the $T$-system recurrence with coefficients to model the generalized determinant, showing it satisfies the Laurent property via cluster algebra structure.
• Construct matrix representations of the $T$-system to derive a determinant formula for the generalized determinant.
• Reformulate the solution in terms of directed networks and path counting, mapping configurations to 6V model states.
• Establish a bijection between ASMs and 6V model configurations with vertex types $a_1, b_1, c_2$ corresponding to entries $0, 0, -1$ in the ASM.
• Prove Theorem 1.2 by transforming the network formulation into the 6V model and deriving the weight assignment per ASM entry.

Experimental results

Research questions

• RQ1 How can the classical Lambda-determinant be generalized to include inhomogeneous parameters $\lambda_a$ and $\mu_a$ along diagonals and anti-diagonals?
• RQ2 What is the closed-form expression for the generalized Lambda-determinant in terms of alternating sign matrices (ASMs)?
• RQ3 How does the generalized determinant relate to the 6V model with domain wall boundary conditions and spectral parameters?
• RQ4 What is the singularity structure of the generating function for the density $\rho_{i,j,k}$ in the inhomogeneous case, and how does it affect the limit shape?
• RQ5 Can the cluster algebra framework be extended to include quantum deformations of the $T$-system and hence of the generalized Lambda-determinant?

Key findings

• The generalized Lambda-determinant is given by a closed formula: $|A|_{\lambda;\mu} = \sum_{B \in \text{ASM}(n)} \prod_{i,j=1}^n w_{i,j}(A,B;\lambda,\mu)$, where weights depend on ASM entries and vertex types.
• For each ASM $B$, the weight $w_{i,j}$ is $a_{i,j}^{b_{i,j}}$ times $\lambda_{j-i}$, $\mu_{i+j-n-1}$, or $\lambda_{j-i} + \mu_{i+j-n-1}$ depending on the 6V vertex type at $(i,j)$, and 1 otherwise.
• In the case $a_{i,j} = 1$ for all $i,j$, the generalized determinant reduces to a sum over ASMs with weights determined solely by $\lambda_a$ and $\mu_b$, yielding a product formula for $n=3$.
• The limit shape of the 6V model under inhomogeneous parameters $\lambda_i = \mu_i = q^i$ is an ellipse: $x^2(1 + \lambda/\mu) + y^2(1 + \mu/\lambda) = 1$, generalizing the Arctic circle theorem.
• The generating function $\rho(X,Y,Z)$ for the density satisfies a non-algebraic functional equation when $\lambda_i = \mu_i = q^i$, with a limit expression involving infinite products and $q$-Pochhammer-like terms.
• The model exhibits a non-trivial singularity structure, and the function $\rho(X,Y,Z)$ is not algebraic for generic $q$, suggesting rich analytic behavior in the inhomogeneous case.