Dice Question Streamline Icon: https://streamlinehq.com

Equality of the right-quiver matrix with the U^∨ factor in Z_P

Prove that, for the partial flag variety G/P, the matrix \tilde{u}_R constructed from the decorated right-quiver Q_{P,R} (as defined in Section 8.2) is equal to the unique matrix u_R \in U^{\vee} that satisfies b_P = u_L \kappa_P \dot{w}_P \bar{w}_0 u_R \in B_-^{\vee}, where u_L is the left-quiver matrix from Section 7.3 and \kappa_P is the highest-weight map from Section 7.2. This establishes that the right-quiver construction recovers precisely the U^{\vee} factor required for b_P to lie in B_-^{\vee}.

Information Square Streamline Icon: https://streamlinehq.com

Background

Section 8 develops a second quiver construction, Q_{P,R}, intended to describe the right-hand factor u_R in the product b_P = u_L \kappa_P \dot{w}_P \bar{w}_0 u_R \in Z_P. The matrix g_R and its U{\vee} factor \tilde{u}_R are defined analogously to g_L and u_L from the left-quiver Q_P.

Conjecture 8.3 asserts that \tilde{u}_R coincides with u_R, thereby giving a fully explicit quiver interpretation of both factors u_L and u_R in Z_P. The authors note that this equality holds for Grassmannians (by Marsh and Rietsch) and expect similar techniques to apply, but they have been unable to complete the proof for general partial flag varieties.

The conjecture is restated equivalently as the identity b_P = g_L \kappa_P \bar{w}_0 \tilde{g}_R, using relationships between Weyl group representatives and the quiver constructions; the authors provide supporting evidence and reductions but indicate that a full proof remains elusive.

References

Conjecture. The matrix \tilde{u}R \in U{\vee} defined in Section \ref{subsec G/P Constructing matrices from the right-quiver decoration} is exactly the matrix u_R \in U{\vee} defined in Section \ref{subsec G/P The quiver torus and toric chart} as the unique matrix such that b_P:= u_L \kappa_P \dot{w}_P \bar{w}_0 u_R lies in B-{\vee}. Due to the similarity between the constructions of g_L and g_R and those used by Marsh and Rietsch, we expect a similar method of proof to work for Conjecture \ref{conj b lies in Z}. However, in our attempts this appears to be much more complicated for partial flag varieties than for Grassmannians, and so at present the conjecture remains open.

Mirror symmetry, tropical geometry and representation theory (2411.04636 - Lüdenbach, 7 Nov 2024) in Conjecture (Section 8.3, "The conjecture")