Parametrizing the Grassmannian using pipe dreams

Kartik Singh · 2025

Read Parametrizing the Grassmannian using pipe dreams by Kartik Singh (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-07627.txt.

Introduction

Kartik Singh reparametrizes the Grassmannian, the space of planes inside larger spaces, using pipe dream diagrams. Postnikov parametrized the totally nonnegative part with networks from special diagrams. Talaska and Williams covered the whole space using Deodhar pieces and Go-diagrams with their own networks.

Pipe dreams tied to those Go-diagrams now supply an alternative network construction. Images of points become easy to compute under the new parametrization. Coordinates that are simple to evaluate help algebraic geometers test positivity and cluster structures. The Grassmannian sits inside scattering amplitude and soliton research, so friendlier coordinates travel far.

Catalog record

CREATORKartik Singh
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE117 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-07627

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Parametrizing the Grassmannian using pipe dreams Kartik Singh Abstract Postnikov gave a parametrization for the totally non-negative Grassmannian using the matroid decomposition and associating a network with L -diagrams. Talaska and Williams extend this result to the entire Grassmannian by using the Deodhar decomposition instead of the matroid decomposition, and the networks this time are associated with the generalized versions of L -diagrams, which are called Go-diagrams. We provide an alternative parametrization for the Deodhar components, this time constructing a network based on the pipe dreams associated with the Go-diagrams. This parametrization has several nice properties; for one, it allows us to easily calculate the image of a point under the isomorphism G ​ r k , n ≃ G ​ r n − k , n Gr_{k,n}\simeq Gr_{n-k,n} . The second feature of this parametrization is that if we write the Plücker coordinates using the Lindstörm-Gessel-Viennot (LGV) lemma in our parametrization, we can associate a pipe dream to each summand, which allows us to reveal additional structure on the summands. Finally, as an application of our parametrization, we describe a case where we can conclude whether one Deodhar component lies inside the closure of another. 1 Introduction The Grassmannian Gr_{k,n} is the set of k -dimensional spaces of \mathbb{F}^{n} where \mathbb{F} is a field. Algebraically we can see it as a coset SL_{n}(\mathbb{F})/P where P is a parabolic subgroup of SL_{n}(\mathbb{F}) . To each point in the Grassmannian, we can assign Plücker coordinates, which can be used to give the Grassmannian a structure of a projective variety. If our field is \mathbb{R} , then the Grassmannian has a very interesting subset called the totally non-negative Grassmannian which is defined as the set of points which have non-negative Plücker coordinates. For a point in the Grassmannian, the set of Plücker coordinates which are non-zero form a basis for a matroid. This allows us to define the matroid strata associated with the matroid \mathcal{M} as the collection of points in the Grassmannian, whose set of non-zero Plücker coordinates forms a basis for \mathcal{M} . This decomposition has a very non-trivial geometry; however, its intersection with the totally non-negative Grassmannian is much more well-behaved. Not all matroids have a non-empty intersection with the totally non-negative Grassmannian; those that do are referred to as positroids. Postnikov showed that the totally non-negative Grassmannian has a CW -complex structure where the cells are given by the positroid components [ Pos06 ] . He further gave a parametrization for each of these positroid cells by assigning a L -diagram to each positroid. Using the L -diagram, he constructed a planar network whose associated weight matrix corresponds to a point in the positroid cell. Talaska and Williams extended this result to obtain a parametrization for the entire Grassmannian [ TW13 ] , where they replaced the positroid cells with Deodhar components, and L -diagrams with Go-diagrams , which are generalizations of L -diagrams. Deodhar components have a simple topology; each component is homeomorphic to (\mathbb{F}^{*})^{a}\times\mathbb{F}^{b} , and we can attach a Go-diagram to each component. The intersection of these with the non-negative Grassmannian is simple to understand as well; the intersection is non-empty if and only if the Go-diagram corresponding to the component is also a L -diagram. In this case, it is the same as the positroid cell corresponding to the L -diagram. Talaska and Williams extended Postnikov’s result by constructing a network for each Go-diagram (which in this case may no longer be planar), such that the weight matrix of this network corresponds to a point in the Deodhar component. When the Go-diagram is a L -diagram, their construction coincides with Postnikov’s network for positroid cells. Deodhar decomposition, even though not as popular as the other decompositions of the Grassmannian, is interesting in its own right. Deodhar first introduced the decomposition in the flag variety to understand the Richardson variety. Kazhdan and Lusztig [ LK79 ] defined R -polynomials to give a recursive formula for the Kazhdan–Lusztig polynomials. R -polynomials themselves can be computed by counting the number of points in the Richardson cell in the setting when our field is finite \mathbb{F}_{q} . Deodhar decomposed each Richardson cell as a union of Deodhar components, and since each Deodhar component has a simple topology, this allowed him to compute the R -polynomials easily. Deodhar decomposition for the Grassmannian was introduced by Kodama and Williams in their work on soliton solutions of the KP equation [ KW13 ] . It is well known that to each point A in the real Grassmannian, we can construct a solition solution u_{A}(x,y,t) of the KP equation. Kodama and Williams show that when t\to\infty , the combinatorics of the solution depend precisely on which Deodhar component A lies in. The Deodhar component, although it has a simple topology and appears to have interesting properties, the decomposition itself is poorly behaved compared to other decompositions. In particular, it is not a stratification (i.e., the closure of a component is not a union of other components) [ Dud08 ] . This makes the problem of understanding the intersection of the closure of one component with another component highly non-trivial. Marcott [ Mar20 ] proposed a conjecture regarding when a component can lie in the closure of another. However, even though his conjecture doesn’t hold in general (see Remark 4.13 ), we were able to show a special case of the conjecture. Both Go-diagrams and L -diagrams have a pipe dream associated with them. In this paper, we will provide an alternative parametrization for the Deodhar components by constructing a network based on these pipe dreams. We will refer to this as the restricted path parametrization . The motivation for introducing these concepts stems from the desire to answer the question: given a Deodhar component with a Go-diagram D , which Deodhar components lie on its codimension-one boundary? Since each component is homeomorphic to (\mathbb{F}^{*})^{a}\times\mathbb{F}^{b} , one way to get something in the codimension one is to take a limit of points with one parameter with values in \mathbb{F}^{*} that is approaching 0. On the level of Go-diagrams, this corresponds to changing filling at a fixed cell or a pair of cells. Therefore, we would like to understand how the parametrization changes when we change the filling at a fixed cell. The changes to the Talaska-Williams network are non-trivial under such changes, and switching to pipe dreams to parametrize these components fixes this problem; each point in the component can be written as a submatrix of a square matrix, which has a product formula that can be easily read off from the pipe dream. There are several advantages of this construction. We have an isomorphism Gr_{k,n}\simeq G_{n-k,n} . If we have a point in Gr_{k,n} along with the parameters corresponding to it, we can write the image of this point under this isomorphism by simply "dualizing" the network on the pipe dreams. This parametrization also gives a new way to look at the Plücker coordinates. Plücker coordinates can be understood using the Lindström-Gessel-Viennot (LGV) lemma on the Talaska–Williams network or the pipe dream network. However, the LGV lemma doesn’t provide any additional structure to the summands that appear. In our restricted path parametrization,

Read the complete text