---
title: Generic Pipe Dream Polynomials
url: https://www.emergentmind.com/topics/generic-pipe-dream-polynomials
type: topic
---

# Generic Pipe Dream Polynomials

Searching arXiv for recent and foundational papers on generic pipe dreams, Schubert/Grothendieck polynomials, and related pipe-dream models.
Generic pipe dream polynomials are weighted generating functions attached to pipe-diagram models that extend the classical role of pipe dreams from monomial formulas for Schubert polynomials to a broader apparatus for equivariant classes, deformations, and interpolating combinatorics. In the strictest sense, the term is introduced for the polynomials
\[
G_w=\sum_{\delta\in GPDs(w)} wt(\delta),
\]
defined by sums over generic pipe dreams and identified with equivariant cohomology classes of lower-upper varieties [2411.11208]. A later hybrid formulation extends the construction to partial permutations \(\pi\), proves independence of the row-by-row hybridization, and shows
\[
G_\pi=(A+B)^m[E_\pi]
\]
for the corresponding component \(E_\pi\) of the lower-upper scheme [2509.01857]. Earlier pipe-dream formulas for double Schubert, Schubert, Grothendieck, and bumpless models supply the immediate antecedents and remain the main comparative context [1909.13777].

## 1. Terminology, scope, and principal frameworks

The terminology is not uniform across the literature. Several foundational papers do not use the exact phrase “generic pipe dream polynomials,” but they provide the closest frameworks: double Schubert pipe dream polynomials as universal equivariant degeneracy-locus classes, Schubert and Grothendieck polynomials as weighted sums over reduced or nonreduced pipe dreams, and bumpless or hybrid models interpolating among different realizations [1909.13777].

| Framework | Indexing data | Polynomial or class |
|---|---|---|
| Double pipe dream polynomial \(C_\pi\) | \(\pi\in S_\infty\) | \(C_\pi=\sum_{D\in PD(\pi)}\prod_{+\in D}(x_{\operatorname{row}(+)}-y_{\operatorname{col}(+)})\) |
| Generic pipe dream polynomial \(G_w\) | \(w\in S_n\) | \(G_w=\sum_{\delta\in GPDs(w)}wt(\delta)\), with \([E_w]=(A+B)^{-n}G_w\) |
| Hybrid generic pipe dream polynomial \(G_\pi\) | \(\pi\in \widetilde S_{m,n}\) | weighted sum over hybrid generic pipe dreams, with \(G_\pi=(A+B)^m[E_\pi]\) |
| Clan polynomial \(s_\gamma(x)\) | \((p,q)\)-clans \(\gamma\) | \(s_\gamma(x)=\sum_{\pi\in BPD(\gamma)}\mathrm{wt}(\pi)\) |

Within this spectrum, two modern constructions are the most literal realizations of the subject. The first is the generic pipe dream polynomial \(G_w\) attached to a permutation \(w\), designed so that classic and bumpless pipe-dream formulas for double Schubert polynomials appear as opposite leading-term regimes [2411.11208]. The second is the hybrid generic pipe dream polynomial \(G_\pi\), attached to a partial permutation \(\pi\), where a row-by-row choice of “classic-like” or “bumpless-like” behavior is shown not to affect the resulting polynomial [2509.01857].

A persistent conceptual point is that “generic” means different things in nearby papers. In the double Schubert setting, it refers to unspecialized equivariant variables \(x_i,y_j\) and universal degeneracy-locus formulas. In the lower-upper-variety setting, it refers to the richer weight system involving \(A\), \(B\), \(x_i\), and \(y_j\), together with generic pipe dream tiles and flux equations. In the Grothendieck setting, a different parameterization is supplied by the deformation parameter \(\beta\) [2407.21653].

## 2. Combinatorial definitions and weight systems

In the construction of Knutson–Zinn-Justin, a generic pipe dream tile is any of seven local tiles, including crossing, elbow or \(j\)-turn, bump, horizontal and vertical straight tiles, reverse elbow, and blank. A generic pipe dream is an \(n\times n\) square tiled by these pieces with boundary conditions: the pipes entering from the west are labeled \(1,\dots,n\), the pipes exiting on the north are labeled \(w^{-1}(1),\dots,w^{-1}(n)\), and the east and south boundary edges are blank. The polynomial is
\[
G_w=\sum_{\delta\in GPDs(w)}wt(\delta),
\]
where
\[
wt(\delta)=\prod_{i,j\in[n]}
\begin{cases}
A+x_i-y_j & \text{if the tile at }(i,j)\text{ is a crossing or a 1-pipe tile},\\[1mm]
B-x_i+y_j & \text{if the tile at }(i,j)\text{ is blank},\\[1mm]
A+B & \text{otherwise.}
\end{cases}
\]
The \(B\)-leading terms come from classic pipe dreams, while the \(A\)-leading terms come from bumpless pipe dreams [2411.11208].

The hybrid theory replaces permutations in \(S_n\) by partial permutations
\[
\pi:\{1,\dots,m\}\to\{1,\dots,n\},
\]
and introduces a hybridization
\[
\beta=(\beta_1,\dots,\beta_m)\in\{\leftarrow,\rightarrow\}^m.
\]
Rows of type \(\leftarrow\) have a pipe entering from the west; rows of type \(\rightarrow\) have a pipe entering from the east; no pipes enter from the south; every pipe exits at the north. Double crossings are allowed. If \(\varphi(i)\) is the label of the pipe entering physical row \(i\), then the tile weights are
\[
A+B
\]
for elbow tiles,
\[
B-x_{\varphi(i)}+y_j,\qquad A+x_{\varphi(i)}-y_j
\]
for blank and straight tiles in west rows, and the same two linear forms with the roles reversed in east rows. The associated polynomial is
\[
G_\pi^\beta=\sum_{\delta:\,\mathrm{conn}(\delta)=\pi}\prod_{(i,j)}wt_{\beta_i}(\delta_{ij},x_{\varphi(i)}-y_j),
\]
and the main theorem states that \(G_\pi^\beta\) is independent of \(\beta\) [2509.01857].

These weight systems differ sharply from the reduced Schubert pipe-dream weights, where a pipe dream contributes a monomial determined solely by row-crossing multiplicities. The 2016 slide-polynomial refinement makes that older situation explicit:
\[
\mathfrak S_w=\sum_{P\in \mathrm{PD}(w)}x^{\mathrm{wt}(P)},
\qquad
\mathrm{wt}(P)_i=\#\{\text{crosses in row }i\},
\]
and then reorganizes the sum by quasi-Yamanouchi pipe dreams and fundamental slide polynomials [1603.09744].

## 3. Relation to Schubert, double Schubert, and Grothendieck families

A central antecedent is the double Schubert pipe dream polynomial
\[
C_\pi=\sum_{D\in \mathrm{PD}(\pi)}\prod_{+\in D}(x_{\operatorname{row}(+)}-y_{\operatorname{col}(+)}),
\]
which is proved to coincide with the algebraically defined double Schubert polynomial and with the equivariant class of the matrix Schubert variety:
\[
A_\pi=C_\pi=G_\pi.
\]
In this sense, the double pipe dream polynomial is already the universal or generic version of the Schubert polynomial, before any specialization in the equivariant parameters [1909.13777].

The newer generic pipe dream polynomials recover the two standard Schubert models by asymptotic or leading-term extraction. For \(G_w\), the highest power of \(B\) comes only from classic pipe dreams and equals
\[
B^{\,n^2-\ell(w)}\,S_w(A+x_1,\dots,A+x_n;\,y_1,\dots,y_n),
\]
while the highest power of \(A\) comes only from bumpless pipe dreams and equals
\[
A^{\,n^2-\ell(w)}\,S_{w_0ww_0}(B-x_n,\dots,B-x_1;\,-y_n,\dots,-y_1).
\]
Accordingly, ordinary and bumpless pipe dreams appear as opposite faces of one larger object [2411.11208]. In the hybrid theory, the same pattern survives in rectangular form: if \(\hat\pi\) is the minimal extension of \(\pi\), then
\[
\operatorname{Lead}_B(G_\pi)=B^{mn-\ell(\hat\pi)}S_\pi(A+x_1,\dots,A+x_m,y_1,\dots,y_n),
\]
and only nongeneric hybrid pipe dreams contribute to that leading form [2509.01857].

Grothendieck polynomials provide a different parameterized pipe-dream family. In that setting,
\[
\mathfrak G_w^\beta(x_1,\ldots,x_n)=\sum_{D\in \mathrm{PD}(w)}\beta^{\#\cross(D)-\ell(w)}\wt(D),
\qquad
\wt(D)=\prod_{(i,j)\in \cross(D)}x_i.
\]
The specialization \(\beta=0\) yields Schubert polynomials, while \(\beta=1\) weights all pipe dreams mapping to \(w\) equally [2407.21653]. This family is not called generic pipe dream polynomials, but it is a natural one-parameter deformation of the same pipe-dream paradigm.

Pipe-dream-based basis changes also extend beyond direct expansions into monomials. Using bumpless pipe dreams and co-BPDs, one has explicit rules for changing bases between Grothendieck and Schubert families:
\[
\mathfrak G_w=\sum_{\substack{\mathcal B\in \operatorname{BPD}(w)\\ \check{\mathcal B}\text{ is reduced}}}
(-1)^{\ell((\check{\mathcal B}))-\ell(w)}\mathfrak S_{(\check{\mathcal B})},
\qquad
\mathfrak S_w=\sum_{\mathcal B\in \operatorname{RBPD}(w)}\mathfrak G_{(\check{\mathcal B})}.
\]
The relevant statistic is the co-permutation of the associated co-BPD, and the canonical bijection of Gao–Huang is shown to preserve co-permutations [2506.07306].

## 4. Geometric interpretations and the meaning of “generic”

The most classical geometric meaning comes from matrix Schubert varieties. In the double Schubert framework, the factor \(x_a-y_b\) attached to a crossing at box \((a,b)\) is the \((T\times T)\)-weight of the matrix coordinate \(m_{ab}\), and the resulting pipe dream polynomial equals the equivariant cohomology class of the matrix Schubert variety \(\overline{X}_\pi\). This makes double pipe dream polynomials universal degeneracy-locus formulas for generic maps of flagged bundles [1909.13777].

The lower-upper-variety theory replaces matrix Schubert varieties by
\[
E:=\{(X,Y): XY \text{ is lower triangular, } YX \text{ is upper triangular}\},
\]
with components \(E_w\). The main class formula is
\[
[E_w]^{B_-\times B_+\times(\mathbb C^\times)^2}=(A+B)^{-n}G_w.
\]
Moreover, \(E_w\) degenerates equivariantly to a union of quadratic complete intersections \(F_\delta\), one for each generic pipe dream \(\delta\), and the class of \(F_\delta\) is exactly the corresponding summand in \((A+B)^{-n}G_w\) [2411.11208].

The hybrid theory extends this to rectangular matrix pairs
\[
E:=\{(X,Y)\in \operatorname{Mat}(m,n)\times\operatorname{Mat}(n,m): XY\text{ lower triangular, }YX\text{ upper triangular}\},
\]
with components \(E_\pi\). Its main theorem is
\[
G_\pi=(A+B)^m[E_\pi].
\]
A second proof degenerates \(E_\pi\) to a union of complete intersections \(V_\delta\) indexed by hybrid generic pipe dreams, with the caveat that embedded components may appear; their absence is conjectured rather than proved [2509.01857].

Bumpless pipe dreams supply a parallel geometric story for diagonal Gröbner degeneration. Under suitable diagonal term orders, the irreducible components of the diagonal Gröbner degeneration of a matrix Schubert variety are indexed by bumpless pipe dreams, counted with scheme-theoretic multiplicity [2108.08370]. Earlier work established this decomposition for classes such as banner permutations and formulated the general diagonal story as dual to the antidiagonal geometry of classical pipe dreams [2003.13719]. One common misconception is therefore misleading: bumpless pipe dreams are not merely an alternative combinatorial expansion, but a geometrically natural model for diagonal degeneration, just as classical pipe dreams are natural for antidiagonal degeneration.

## 5. Structural mechanisms: Yang–Baxter, flux, and canonical compression

Hybrid generic pipe dream polynomials are controlled by two structural principles. The first is Yang–Baxter symmetry: adjacent row types can be swapped without changing the polynomial, and the resulting state sums satisfy a divided-difference recurrence. If \(\pi'=\pi r_i\), then
\[
G_{\pi} = \frac{1}{x_i-x_{i+1}}
\left( (A+B)\,G_{\pi'} - (A+B+x_i-x_{i+1})\,r_iG_{\pi'} \right).
\]
For decreasing \(\pi\), there is also an explicit product formula. These recurrences are proved by inserting an auxiliary diamond and moving it across the diagram by Yang–Baxter moves [2509.01857].

The second principle is flux. Edge fluxes \(\Phi_e\) are defined from the matrix entries, and local conservation laws imply that pipes may be recovered from equalities among fluxes. The component \(E_\pi\) is characterized by flux equations
\[
(YX)_{jj}=
\begin{cases}
(XY)_{ii} & \text{if } j=\pi(i),\\
0 & \text{otherwise,}
\end{cases}
\]
so connectivity becomes a statement about which flux exits at which top edge. This is a distinctive feature of the generic setting: the notion of pipe dream is derived from equalities among fluxes rather than postulated independently [2509.01857].

A different but closely related structural theme appears in the slide-polynomial refinement of Schubert combinatorics. There one compresses the full pipe-dream sum by destandardization to quasi-Yamanouchi pipe dreams and obtains
\[
\mathfrak S_w=\sum_{P\in \mathrm{QPD}(w)}\mathfrak F_{\mathrm{wt}(P)}.
\]
This packages the monomial generating function of all reduced pipe dreams into a smaller canonical indexing set and refines the stabilization from Schubert polynomials to Stanley symmetric functions [1603.09744]. In another direction, reduced pipe dreams carry a Demazure crystal structure via crystal chute moves, and this yields a decomposition of Schubert polynomials into key polynomials indexed by highest-weight pipe dreams [2403.07204]. These constructions do not define generic pipe dream polynomials, but they show that pipe-dream-generated polynomials admit canonical quotients, internal crystals, and basis refinements.

## 6. Extensions beyond the basic Schubert setting

Several later developments broaden the range of pipe-dream-type polynomial theories. For \((p,q)\)-clans, bumpless pipe dream fragments on a clan-dependent Young diagram define clan polynomials
\[
s_\gamma(x)=\sum_{\pi\in BPD(\gamma)}\mathrm{wt}(\pi),
\]
and these coefficients appear exactly in the equivariant Schubert expansion of the \(GL_p\times GL_q\)-orbit closure class \([Y_\gamma]_T\). In the full-rectangle case, the clan polynomial reduces to a double Schubert polynomial of an associated partial permutation [2511.00980].

Quantum deformation is realized by quantum bumpless pipe dreams. In that model,
\[
S_w^q(x,y)=\sum_{P\in \QBPD(w)}\bwt(P),
\]
where \(\bwt(P)\) includes factors \(x_i-y_j\), \(q_i\), and \(-q_i\) depending on empty, domino, upward-cross, southwest-elbow, and upward-vertical tiles. The formula specializes to ordinary bumpless pipe dreams at \(q=0\), but it is not cancellation-free; the paper gives explicit examples of complete and partial cancellation [2403.16168]. This resolves a common expectation inherited from cohomological formulas: in quantum settings, a pipe-dream model may exist even when positivity does not.

Pipe puzzles provide another hybrid extension. For permutations \(u,v\in S_n\) with separated descents at position \(k\), the coefficients in
\[
\mathfrak G_u(x,y)\,\mathfrak G_v(x,t)
=
\sum_w c^w_{u,v}(t,y)\,\mathfrak G_w(x,t)
\]
are given by
\[
c^w_{u,v}(t,y)=\sum_{\pi\in \mathrm{PP}(u,v,w)}\mathrm{wt}(\pi).
\]
This model incorporates both bumpless pipe-dream and puzzle structures, specializes to the Knutson–Zinn-Justin separated-descent rule when \(y=t\), and recovers Weigandt’s bumpless formula when \(v=\mathrm{id}\) and \(x=t\) [2309.00467].

Grothendieck top-degree theory supplies yet another specialized variant. For inverse fireworks permutations, the top homogeneous component \(\widehat{\mathfrak G}_w(x)\) admits the direct formula
\[
\widehat{\mathfrak G}_w(x)=\sum_{B\in BVPD(w)} wt(B),
\]
where \(BVPD(w)\) denotes bumpless vertical-less pipe dreams. This gives the first direct combinatorial formula for the Castelnuovo–Mumford polynomial in that regime [2403.12186]. At the opposite scale, principal specializations of \(\beta\)-Grothendieck polynomials support a probabilistic model of random permutations from pipe dreams, with the \(\beta=1\) case occupying the paper’s main asymptotic focus [2407.21653].

Taken together, these constructions show that generic pipe dream polynomials are not a single isolated family but a node in a larger theory of weighted pipe-dream state sums. Across Schubert, Grothendieck, lower-upper, clan, quantum, and puzzle settings, the recurring pattern is a locally defined pipe model whose global generating function represents a canonical geometric or representation-theoretic object.

Source: https://www.emergentmind.com/topics/generic-pipe-dream-polynomials