Plücker relations for quiver Grassmannians (Q1737225): Difference between revisions
From MaRDI portal
Set OpenAlex properties. |
Changed an Item |
||
Property / arXiv ID | |||
Property / arXiv ID: 1607.01058 / rank | |||
Normal rank |
Revision as of 21:33, 18 April 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Plücker relations for quiver Grassmannians |
scientific article |
Statements
Plücker relations for quiver Grassmannians (English)
0 references
27 March 2019
0 references
Let $Q=(Q_0,Q_1)$ be a quiver, and let $M$ be a finite-dimensional complex representation of $Q$, and $\underline{d}=\underline{\dim} (M)$, the dimension vector of $M$. Let $\underline{e}=(e_p)_{p\in Q_0}$, a dimension vector. The quiver Grassmannian $\text{Gr}_{\underline{e}}(M)$ is the collection of all subrepresentations $N$ of $M$ with dimension vector $\underline{\dim} (N)=\underline{e}$. Furthermore, it can be described as the closed embedding \[ \iota: \text{Gr}_{\underline{e}}(M) \hookrightarrow \prod_{p\in Q_0} \text{Gr}(e_p,M_p) \] into the product of usual Grassmannians $\text{Gr}(e_p,M_p)$. The choice of ordered bases $\mathcal{B}_p$ of $M_p$ for each $p\in Q_0$ identifies $M_p\cong \mathbb{C}^{d_p}$, and yields Plücker coordinates $[\ldots: \Delta_I:\ldots]$ for the complex points of $\text{Gr}(e_p,M_p)$, where $I$ varies over all cardinality $e_p$-subsets of $\mathcal{B}_p$. \par Next, quiver Plücker relations are described as follows. Let $m_{v,j,i}$ be the matrix coefficients of the linear maps $M_v:M_p\to M_q$. For a subset $I\cup\{i\}$ of $\mathcal{B}_p$, define $\epsilon(i,I)=\#\{i'\in I : i'\leq i\}$. Then $\text{im}(\text{Gr}_{\underline{e}}(M)\hookrightarrow \prod_{p\in Q_0}\text{Gr}(e_p,M_p))$ is carved out by quiver Plücker relations \[ E(v,I,J) = \sum_{i\in \mathcal{B}_p-I,\, j\in J} (-1)^{\epsilon(i,I)+\epsilon(j,J)} m_{v,j,i} \Delta_{I\cup\{i\}} \Delta_{J-\{j\}} = 0, \] where $v:p\to q$ varies over the arrows $Q_1$ of $Q$, $I$ varies over the $(e_p-1)$-subsets of $\mathcal{B}_p$, and $J$ varies over the $(e_q+1)$-subsets of $\mathcal{B}_q$. \par Define $\text{Rep}(\underline{d}) := \prod_{v:p\to q\text{ in }Q} \text{Hom}(\mathbb{C}^{d_p},\mathbb{C}^{d_q})$, the moduli space of $\underline{d}$-dimensional representations of $Q$ with fixed basis. The complex points of $\text{Rep}(\underline{d})$ correspond to tuples of matrices $A_v\in\text{Hom}(\mathbb{C}^{d_p},\mathbb{C}^{d_q})$ indexed by the arrows $v:p\to q$, where $v\in Q_1$. Consider the trivial fiber bundle \[ \pi: \text{Rep}(\underline{d})\, \times\prod_{p\in Q_0}\text{Gr}(e_p,d_p) \ \longrightarrow \ \text{Rep}(\underline{d}) \] whose complex points correspond to pairs of a tuple $(A_v)_{v\in Q_1}$ of matrices $A_v\in\text{Hom}(\mathbb{C}^{d_p},\mathbb{C}^{d_q})$ and a tuple $(N_p)_{p\in Q_0}$ of $e_p$-dimensional subspaces $N_p \subseteq \mathbb{C}^{d_p}$. \par The universal quiver Grassmannian is the closed subscheme $\text{Gr}_{\underline{e}}^Q(\underline{d})\subseteq \text{Rep}(\underline{d})\times\prod_p\text{Gr}(e_p,d_p)$ that is defined by the conditions $A_v(N_p) \subset N_q$ for every arrow $v:p\to q$, $v\in Q_1$. Let $\pi_Q := \pi|_{\text{Gr}_{\underline{e}}^Q(\underline{d})}$. The choice of ordered bases $\mathcal{B}_p$ of $M_p$ defines matrices $A_v=(m_{v,j,i})_{i\in\mathcal{B}_p, j\in\mathcal{B}_q}$, and thus a point $(A_v)_{v\in Q_1}\in \text{Rep}(\underline{d})$. As a scheme, the quiver Grassmannian $\text{Gr}_{\underline{e}}(M)$ is the fiber of $\pi_Q$ over $(A_v)_{v\in Q_1}$, and the embedding of $\text{Gr}_{\underline{e}}(M)$ into the fiber of $\pi$ over $(A_v)_{v\in Q_1}$ coincides with the embedding into $\prod_{p\in Q_0}\text{Gr}(e_p,d_p)$. \par The authors study quiver Plücker relations for $\text{Gr}_{\underline{e}}(M)$ by describing it as a closed subscheme of a product of usual Grassmannians (Theorem 2.2, page 213): \[ \text{im}\left( \text{Gr}_{\underline{e}}(M)\to \prod_{p\in Q_0}\text{Gr}(e_p,d_p)\right) \] is the subscheme defined by the quiver Plücker relations \[ E(v,I,J) \ = \ \sum_{i\in \mathcal{B}_p-I,\, j\in J} (-1)^{\epsilon(i,I)+\epsilon(j,J)} \, m_{v,j,i} \, \, \Delta_{I\cup\{i\}} \Delta_{J-\{j\}} \ = \ 0 \] where $v:p\to q$ ranges over all arrows of $Q$, $I$ ranges over all $(e_p-1)$-subsets of $\mathcal{B}_p$, and $J$ ranges over all $(e_q+1)$-subsets of $\mathcal{B}_q$.
0 references
quiver Grassmannians
0 references
Plücker relations
0 references