ScalingStacks

6.1. Quiver flag varieties

We first recall some basic definitions and facts about quiver flag varietes. We refer to [SW14] and [Prz19] for more details. Consider a quiver QQ with finite sets of vertices Q0Q_{0} and arrows Q1Q_{1} and source and target maps

s,t:Q1⇉Q0.s,t:Q_{1}\rightrightarrows Q_{0}.

Denote by Γ=ℤ≥0​Q0\Gamma=\mathbb{Z}_{\geq 0}Q_{0} and Γ+=Γ\{0}\Gamma^{+}=\Gamma\backslash\{0\} the sets of (non-trivial) dimension vectors. The dimension vector of a Q0Q_{0}-graded kk-vector space V=(Vi)i∈Q0V=(V_{i})_{i\in Q_{0}} is dim(V)=(dim(Vi))i∈Q0∈Γ.\dim(V)=(\dim(V_{i}))_{i\in Q_{0}}\in\Gamma. We denote by

Rep⁡(V)=∏a∈Q1Homk⁡(Vs⁡(a),Vt⁡(a))\operatorname{Rep}(V)=\prod_{a\in Q_{1}}\operatorname{Hom}_{k}(V_{s(a)},V_{t(a)})

the vector space of quiver representations with underlying Q0Q_{0}-graded vector space V.V. We fix a dimension vector 𝐝∈Γ\mathbf{d}\in\Gamma and let VV be the standard vector space with dimV=𝐝.\dim V=\mathbf{d}. We often abbreviate Rep⁡(𝐝)=Rep⁡(V).\operatorname{Rep}(\mathbf{d})=\operatorname{Rep}(V). There is a a natural conjugation action on Rep⁡(𝐝)\operatorname{Rep}(\mathbf{d}) by the group

GL⁡(𝐝)=∏i∈Q0GL⁡(Vi)=∏i∈Q0GL𝐝i⁡(k).\operatorname{GL}(\mathbf{d})=\prod_{i\in Q_{0}}\operatorname{GL}(V_{i})=\prod_{i\in Q_{0}}\operatorname{GL}_{\mathbf{d}_{i}}(k).

A composition of a dimension vector 𝐝\mathbf{d} is a tuple 𝐝¯=(𝐝¯j)∈(Γ+)ℓ𝐝¯\underline{\mathbf{d}}=(\underline{\mathbf{d}}^{j})\in(\Gamma^{+})^{\ell_{\underline{\mathbf{d}}}} which sums to 𝐝.\mathbf{d}. A composition 𝐝¯\underline{\mathbf{d}} is called complete if each 𝐝¯j\underline{\mathbf{d}}^{j} is a unit vector. We write Comp⁡(𝐝)\operatorname{Comp}(\mathbf{d}) and Compf⁡(𝐝)\operatorname{Compf}(\mathbf{d}) for the set of (complete) compositions of 𝐝.\mathbf{d}. A partial flag V¯\underline{V} of VV of type 𝐝¯\underline{\mathbf{d}} is a sequence of Q0Q_{0}-graded kk-vector spaces

0=V0⊂V1⊂⋯⊂Vℓ𝐝¯=V0=V^{0}\subset V^{1}\subset\dots\subset V^{\ell_{\underline{\mathbf{d}}}}=V

such that dimVj/Vj−1=𝐝¯i.\dim V^{j}/V^{j-1}=\underline{\mathbf{d}}^{i}. We denote the smooth projective variety of such flags by Fl⁡(V,𝐝¯)=Fl⁡(𝐝¯).\operatorname{Fl}(V,\underline{\mathbf{d}})=\operatorname{Fl}(\underline{\mathbf{d}}). The partial flag V¯\underline{V} is called strictly ρ\rho-stable for a quiver representation ρ∈Rep⁡(𝐝)\rho\in\operatorname{Rep}(\mathbf{d}) if

(6.1) ρa​(Vs⁡(a)j)⊂Vt⁡(a)j−1​ for all ​i=1,…,ℓ𝐝¯​ and ​a∈Q1.\displaystyle\rho_{a}(V^{j}_{s(a)})\subset V^{j-1}_{t(a)}\text{ for all }i=1,\dots,\ell_{\underline{\mathbf{d}}}\text{ and }a\in Q_{1}.

We can hence consider the variety

𝔔⁡(𝐝¯)={(ρ,V¯)∈Rep⁡(𝐝)×Fl⁡(𝐝¯)∣V¯​ is strictly ρ-stable}.\mathfrak{Q}(\underline{\mathbf{d}})=\{(\rho,\underline{V})\in\operatorname{Rep}(\mathbf{d})\times\operatorname{Fl}(\underline{\mathbf{d}})\mid\underline{V}\text{ is strictly $\rho$-stable}\}.

The variety 𝔔⁡(𝐝¯)\mathfrak{Q}(\underline{\mathbf{d}}) is smooth and has a diagonal action by GL⁡(𝐝).\operatorname{GL}(\mathbf{d}). Projection yields a GL⁡(𝐝)\operatorname{GL}(\mathbf{d})-equivariant proper map

μ𝐝¯:𝔔⁡(𝐝¯)→Rep⁡(𝐝).\mu_{\underline{\mathbf{d}}}:\mathfrak{Q}(\underline{\mathbf{d}})\to\operatorname{Rep}(\mathbf{d}).

For a representation M=(V,ρ)M=(V,\rho), the fiber μ𝐝¯−1​(ρ)\mu_{\underline{\mathbf{d}}}^{-1}(\rho) is called a (partial) quiver flag variety and denoted by

Fl⁡(M,𝐝¯)={V¯∈Fl⁡(V,𝐝¯)∣V¯​ is strictly ρ-stable}.\operatorname{Fl}(M,\underline{\mathbf{d}})=\{\underline{V}\in\operatorname{Fl}(V,\underline{\mathbf{d}})\mid\underline{V}\text{ is strictly $\rho$-stable}\}.

Original mathematics by the credited authors. Source collection and HTML conversion remain in progress.

Jens Niklas Eberhardt, Catharina Stroppel

Original source: arXiv:2109.00305v2