betaFeedback

WorksMathematics

Stationary Measures and Boundary Correlations for Multispecies Half-Line ASEP

Abstract

We construct stationary measures of the multispecies ASEP on the positive half-line, with injection of highest-priority particles at the left boundary, using vertex models. Our results extend existing constructions of Martin and Aggarwal-Nicoletti-Petrov for the full-line ASEP to the half-line. For the two-species process, we compute the joint distribution of occupation colors at the first two sites. We also partially address the connection with the half-space ASEP speed process.

Introduction

Overview

The asymmetric simple exclusion process (ASEP) is a basic model of nonequilibrium transport, and at the same time one of the principal exactly solvable interacting particle systems. Particles attempt to exchange positions with neighboring holes at unequal rates in the two directions. The asymmetry creates a nonzero stationary current and, on large scales, connects exclusion dynamics to the theory of hyperbolic conservation laws and to the Kardar-Parisi-Zhang (KPZ) universality class. On the infinite integer line, the single-species ASEP has simple translation-invariant Bernoulli stationary measures; introducing an open boundary, however, leads to substantially richer stationary laws. In particular, exact phase diagrams, density profiles, and correlation functions of the open ASEP can be accessed by the matrix-product ansatz [74] and its subsequent refinements. We refer to [91, 83, 84, 76, 59, 92, 66, 65, 58, 93] for a non-exhaustive list of key works on single-species ASEP stationary measures.

In a multispecies exclusion process, every site carries one of several particle colors or a hole, and neighboring colors exchange according to a fixed priority order. Merging the higher-priority colors produces an ordinary one-species ASEP, but the collection of these projections does not determine the correlations of the full colored stationary measure. Already with two species, the stationary distributions exhibit nontrivial clustering and microscopic shock structure [75, 77, 52]. On the infinite integer line, stationary measures for arbitrary numbers of species in the totally asymmetric case were constructed using multiline diagrams and queues [78]. The partially asymmetric case was treated using a matrix-product representation [88] and a queueing formulation [86]; see also the semi-infinite single-species matrix-product approach in [89].

On the integer line, queueing constructions are especially useful for explicit computations with stationary distributions. They have led, for instance, to beautiful formulas for the TASEP speed process and multispecies correlations [51, 53, 54]. Almost-sure existence of the full-line ASEP speed process was proved in [48]; related half-line second-class particle asymptotics have been studied in [69]. More recently, such constructions were interpreted in the language of colored stochastic vertex models [50], with the Yang-Baxter equation playing a central role. Relevant developments in stochastic higher-spin and colored vertex models include [70, 61, 63, 60, 68].

A boundary changes the stationary measures of multispecies ASEP in an essential way. Nevertheless, the vertex-model connection suggests a possibility of a half-line vertex model construction: at an integrable boundary the Yang-Baxter equation is supplemented by the reflection equation [90]. Related half-space integrable models and half-line ASEP have been studied in [56, 81, 82, 57]. Integrable boundary weights and half-space six-vertex constructions appear in [85, 79]; see also the recent half-space TASEP transition formulas [73] and work on stationary measures for six-vertex models on a strip [94].

The present paper provides such a vertex construction. We study the positive half-line ASEP, with particles of the strongest color injected at an arbitrary rate at the left endpoint. Our first main result, Theorem 1.2, is a vertex-model construction of stationary measures for the multispecies process. Our second main result, Theorem 1.4, gives an explicit calculation of the joint occupation distribution at the first two sites of the half-line; obtaining this formula was our original motivation. As explained below, this distribution is also related to the joint distribution of speeds in the half-space ASEP speed process, conditional on plausible conjectures that we do not prove here.

It should be noted that we study only a special case of multispecies ASEP on the half-line. Our boundary update is totally asymmetric, and only first-class particles can be injected. There are two reasons for the restriction. First, it is this version of ASEP, with the injection of first-class particles only, that appears in the description of speeds in the half-space ASEP speed process. This makes this part of the restriction more natural. Second, while it is plausible that the general strategy for constructing stationary measures works for a wider class of boundary interactions, the technical details become more complicated, and we prefer to avoid them.

During our work on this project, we learned of independent related work in progress [49].

Let us present the results in more detail.

The stationary measure construction

First, let us define in more detail the version of the multispecies ASEP studied in this paper. Fix the number of species r∈Z≥1r\in\mathbb Z_{\ge1}. The state space is {0,1,…,r}Z≥1\{0,1,\ldots,r\}^{\mathbb Z_{\ge 1}}, and the priority order is 1≻2≻⋯≻r≻01\succ2\succ\cdots\succ r\succ0; color 00 is a hole. A stronger color (we use the words species, color, type and class interchangeably) exchanges with a weaker color to its right at rate 11 and to its left at rate q∈[0,1)q\in[0,1). At site 11, every color other than 11 is replaced by 11 at rate (1−q)α(1-q)\alpha, where α>0\alpha>0.

Fix density parameters ρ1,…,ρr\rho_1,\ldots,\rho_r such that

0<ρ1<⋯<ρr<1,ρ1>max⁡(1/2,1−α).\begin{equation} 0<\rho_1<\cdots<\rho_r<1,\qquad \rho_1>\max(1/2,1-\alpha). \tag*{(1.1)} \end{equation}

Our goal is to construct a stationary measure μα,ρ1,…,ρr\mu^{\alpha,\rho_1,\ldots,\rho_r} of ASEP, which has asymptotic density ρ1\rho_1 of first-class (or color 1) particles, asymptotic density ρ2\rho_2 of color 1 and color 2 particles, ..., asymptotic density ρr\rho_r of colors from 11 to rr particles. Consequently, the density of holes (color 0) is 1−ρr1-\rho_r.

In short, this stationary measure can be read off from the right-hand side of Figure 4 below, in which we apply to trivial input from the left and from below a collection of stochastic vertices (also, the limit ”height to infinity” of this picture has to be taken in order to get half-space), which produces a probability measure on the right. Let us define all necessary ingredients.

We construct a stationary measure from three types of stochastic vertices. To distinguish various components of these vertices more easily, we introduce the following terminology. A physical edge carries a single color (it corresponds eventually to the space of ASEP). A carrier edge contains infinitely many colors and records the numbers of arrows of the relevant colors. It corresponds to queues. Write Cm={0,1,…,m}\mathcal C_m=\{0,1,\ldots,m\} for 0≤m≤r0\le m\le r, and set x=max⁡{0,1−α−1}x=\max\{0,1-\alpha^{-1}\}.

Usual vertices.

Fix 1≤m≤r1\le m\le r. The usual vertex W(m)(G,i;G′,j)\mathbb W^{(m)}(G,i;G',j) has incoming carrier GG and physical color i∈Cm−1i\in\mathcal C_{m-1}, and outgoing carrier G′G' and color j∈Cmj\in\mathcal C_m. Its carrier states have the form

G=(G1,…,Gm−1,∞),G1,…,Gm−1∈Z≥0.G=(G_1,\ldots,G_{m-1},\infty),\qquad G_1,\ldots,G_{m-1}\in\mathbb Z_{\ge0}.

The last coordinate represents an unlimited amount of color mm. Adding or subtracting EmE_m does not change this coordinate. Here EkE_k is the standard basis vector in coordinate kk, and G[1,k−1]=∑a=1k−1GaG_{[1,k-1]}=\sum_{a=1}^{k-1}G_a. The nonzero weights are displayed in Figure 1. The carrier is read from bottom to top and the physical color from left to right.

tikzpicture[baseline=0pt, every node/.style=font=,inner sep=1pt] [paper normal,line width=3.8pt, -Stealth[length=5.5pt,width=7.5pt]] (0,-.60)--(0,.70); [paper normal,line width=1.5pt, -Stealth[length=4pt,width=4.5pt]] (-.72,0)--(.78,0); [above=3pt] at (0,.70)GG; [left=4pt] at (-.72,0)00; [right=4pt] at (.78,0)00; [below=3pt] at (0,-.60)GG; tikzpicturetikzpicture[baseline=0pt, every node/.style=font=,inner sep=1pt] [paper normal,line width=3.8pt, -Stealth[length=5.5pt,width=7.5pt]] (0,-.60)--(0,.70); [paper normal,line width=1.5pt, -Stealth[length=4pt,width=4.5pt]] (-.72,0)--(.78,0); [above=3pt] at (0,.70)GG; [left=4pt] at (-.72,0)kk; [right=4pt] at (.78,0)kk; [below=3pt] at (0,-.60)GG; tikzpicturetikzpicture[baseline=0pt, every node/.style=font=,inner sep=1pt] [paper normal,line width=3.8pt, -Stealth[length=5.5pt,width=7.5pt]] (0,-.60)--(0,.70); [paper normal,line width=1.5pt, -Stealth[length=4pt,width=4.5pt]] (-.72,0)--(.78,0); [above=3pt] at (0,.70)G−EkG-E_k; [left=4pt] at (-.72,0)00; [right=4pt] at (.78,0)kk; [below=3pt] at (0,-.60)GG; tikzpicture
1−ρm\displaystyle 1-\rho_mρmqG[1,k−1](1−xqGk)\displaystyle \rho_m q^{G_{[1,k-1]}}(1-x q^{G_k})ρmqG[1,k−1](1−qGk)\displaystyle \rho_m q^{G_{[1,k-1]}}(1-q^{G_k})
tikzpicture[baseline=0pt, every node/.style=font=,inner sep=1pt] [paper normal,line width=3.8pt, -Stealth[length=5.5pt,width=7.5pt]] (0,-.60)--(0,.70); [paper normal,line width=1.5pt, -Stealth[length=4pt,width=4.5pt]] (-.72,0)--(.78,0); [above=3pt] at (0,.70)G+EkG+E_k; [left=4pt] at (-.72,0)kk; [right=4pt] at (.78,0)00; [below=3pt] at (0,-.60)GG; tikzpicturetikzpicture[baseline=0pt, every node/.style=font=,inner sep=1pt] [paper normal,line width=3.8pt, -Stealth[length=5.5pt,width=7.5pt]] (0,-.60)--(0,.70); [paper normal,line width=1.5pt, -Stealth[length=4pt,width=4.5pt]] (-.72,0)--(.78,0); [above=3pt] at (0,.70)G+Ek−EℓG+E_k-E_\ell; [left=4pt] at (-.72,0)kk; [right=4pt] at (.78,0)ℓ\ell; [below=3pt] at (0,-.60)GG; tikzpicturetikzpicture[baseline=0pt, every node/.style=font=,inner sep=1pt] [paper normal,line width=3.8pt, -Stealth[length=5.5pt,width=7.5pt]] (0,-.60)--(0,.70); [paper normal,line width=1.5pt, -Stealth[length=4pt,width=4.5pt]] (-.72,0)--(.78,0); [above=3pt] at (0,.70)G+Eℓ−EkG+E_\ell-E_k; [left=4pt] at (-.72,0)ℓ\ell; [right=4pt] at (.78,0)kk; [below=3pt] at (0,-.60)GG; tikzpicture
1−ρm\displaystyle 1-\rho_mρmxqG[1,ℓ−1](1−qGℓ)\displaystyle \rho_m x q^{G_{[1,\ell-1]}}(1-q^{G_\ell})ρmqG[1,k−1](1−qGk)\displaystyle \rho_m q^{G_{[1,k-1]}}(1-q^{G_k})

Figure 1. The usual vertex W(m)\mathbb W^{(m)}. The carrier edge is read from bottom to top and the physical edge from left to right. Thick gray edges carry carrier states and thin gray edges carry physical colors. We have 1≤k<ℓ≤m1\le k<\ell\le m.

Return vertices.

The return vertex L^ver(i,H;j,H′)\widehat L^{\mathrm{ver}}(i,H;j,H') has incoming physical color ii and carrier HH, and outgoing color jj and carrier H′H', with i,j∈Cmi,j\in\mathcal C_m. The carrier edge should be viewed as almost entirely filled with color 11, with finitely many ”defects” of colors 0,2,…,m0,2,\ldots,m. It is indexed via

C^m=Cm∖{1},H=(H0,H2,…,Hm)∈Z≥0C^m.\widehat{\mathcal C}_m=\mathcal C_m\setminus\{1\},\qquad H=(H_0,H_2,\ldots,H_m) \in\mathbb Z_{\ge0}^{\widehat{\mathcal C}_m}.

Color 11 is the implicit background and is not recorded in HH; a physical input 11 therefore adds no carrier occupation. Write ∣H∣=H0+∑a=2mHa|H|=H_0+\sum_{a=2}^m H_a, and let E^k\widehat E_k denote the standard basis vector in coordinate k∈C^mk\in\widehat{\mathcal C}_m. Figure 2 gives all nonzero weights. The vertex weights satisfy H′=H+E^i−E^jH'=H+\widehat E_i-\widehat E_j, where E^1=0\widehat E_1=0. The physical color is read from left to right, as before, but the return carrier is read from top to bottom. For example, in the bottom-right configuration, the carrier coming from the top receives color ℓ\ell from the left and releases color kk to the right.

Both usual and return vertices have a useful queue-type description, see Proposition 4.11 below.

References

  1. [48]ACG23. (2023). The ASEP speed process,.
  2. [49]A. Aggarwal, I. Corwin, J. He, M. Hegde, and M. Wheeler, \newblock In preparation.
  3. [50]ANP25. (2025). Colored interacting particle systems on the ring: stationary measures from Yang--Baxter equations,.
  4. [51]AAV11. (2011). The TASEP speed process,.
  5. [52]Ang06. (2006). The stationary measure of a 2-type totally asymmetric exclusion process,.
  6. [53]AL14. (2014). An inhomogeneous multispecies TASEP on a ring,.
  7. [54]AL17. (2017). Correlations in the multispecies TASEP and a conjecture by Lam ,.
  8. [55]Bah12. (2012). Hydrodynamics and hydrostatics for a class of asymmetric particle systems with open boundaries,.
  9. [56]BBCW18. (2018). Stochastic six-vertex model in a half-quadrant and half-line open ASEP,.
  10. [57]BC24. (2024). Markov duality and Bethe ansatz formula for half-line open ASEP,.
  11. [58]BLD23. (2023). Stationary measures of the KPZ equation on an interval from Enaud--Derrida 's matrix product ansatz representation,.
  12. [59]BE07. (2007). Nonequilibrium steady states of matrix-product form: a solver's guide,.
  13. [60]BCG16. (2016). Stochastic six-vertex model,.
  14. [61]BP18. (2018). Higher spin six vertex model and symmetric rational functions,.
  15. [62]BW21. (2021). Spin $q$-Whittaker polynomials,.
  16. [63]BW22. (2022). Colored stochastic vertex models and their spectral theory,.
  17. [64]BM16. (2016). Construction of $R$ -matrices for symmetric tensor representations related to $U_q( sl _n)$ ,.
  18. [65]BWW23. (2023). From the asymmetric simple exclusion processes to the stationary measures of the KPZ fixed point on an interval,.
  19. [66]BW17. (2017). Asymmetric simple exclusion process with open boundaries and quadratic harnesses,.
  20. [67]Buf20. (2003). Interacting particle systems and random walks on Hecke algebras,. https://arxiv.org/abs/2003.02730
  21. [68]BMP21. (2021). Yang--Baxter random fields and stochastic vertex models,.
  22. [69]Che26. (2026). The second-class particle in the half-line open totally asymmetric simple exclusion process,.
  23. [70]CP16. (2016). Stochastic higher spin vertex models on the line,.
  24. [71]CEMRV16. (2016). Matrix product solution to a 2-species TASEP with open integrable boundaries,.
  25. [72]DZ26. (2026). The directed landscape in half-space,. https://arxiv.org/abs/2604.10020
  26. [73]dGMRW25. (2025). Pfaffian transition probability and correlation kernel of TASEP in half-space,. https://arxiv.org/abs/2512.02286
  27. [74]DEHP93. (1993). Exact solution of a 1D asymmetric exclusion model using a matrix formulation,.
  28. [75]DJLS93. (1993). Exact solution of the totally asymmetric simple exclusion process: shock profiles,.
  29. [76]DLS03. (2003). Exact large deviation functional of a stationary open driven diffusive system: the asymmetric exclusion process,.
  30. [77]FFK94. (1994). Invariant measures for a two-species asymmetric process,.
  31. [78]FM07. (2007). Stationary distributions of multi-type totally asymmetric exclusion processes,.
  32. [79]GdGMW25. (2025). Symmetric functions from the six-vertex model in half-space,.
  33. [80]Gro04. (2004). Phase Transitions in Nonequilibrium Stochastic Particle Systems with Local Conservation Laws,.
  34. [81]He24. (2024). Boundary current fluctuations for the half-space ASEP and six-vertex model,.
  35. [82]He26. (2026). Shift invariance of half space integrable models,.
  36. [83]Lig75. (n.d.). Ergodic theorems for the asymmetric simple exclusion process,.
  37. [84]Lig85. (1985). Interacting Particle Systems,.
  38. [85]ML19. (2019). Boundary matrices for the higher spin six vertex model,.
  39. [86]Mar20. (2020). Stationary distributions of the multi-type ASEP ,.
  40. [87]PdS25. (2025). Correlations in the continuous multispecies TASEP on a ring,.
  41. [88]PEM09. (2009). The matrix product solution of the multispecies partially asymmetric exclusion process,. https://arxiv.org/abs/0812.3293
  42. [89]SasWil14. (2014). Combinatorics of the asymmetric exclusion process on a semi-infinite lattice,.
  43. [90]Sk88. (1988). Boundary conditions for integrable quantum systems,.
  44. [91]Sp70. (n.d.). Interaction of Markov processes,.
  45. [92]USW04. (2004). Asymmetric simple exclusion process with open boundaries and Askey--Wilson polynomials,.
  46. [93]WWY24. (2024). Askey--Wilson signed measures and open ASEP in the shock region,.
  47. [94]Yan24. (2024). Stationary measure for six-vertex model on a strip,.

Paper details

vlnc:W_w7rz7ux2
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

ChatGPT

View statement
Looking for related papers…
Contents