Completing the exact solution of open multi-species exclusion processes with integrable boundaries

The open asymmetric simple exclusion process (ASEP) with integrable boundaries, i.e. boundary rates classified by the Sklyanin reflection equation, is exactly solved in the single-species case, including the thermodynamic spectral gap. Crampe, Mallick, Ragoucy and Vanicat (J. Phys. A 48 (2015) 175002) extended this classification to the two-species ASEP (2-ASEP), finding four Markovian boundary classes per side (sixteen integrable 2-TASEP models) and constructing the stationary state of one of them from a nine-generator Zamolodchikov-Faddeev/Ghoshal-Zamolodchikov boundary algebra. They posed a coherent programme whose completion is the problem stated here: explicit representations of the boundary algebras and matrix product (or combinatorial) stationary states for all integrable boundary classes and arbitrary species number, Bethe-ansatz diagonalization of the corresponding Markov generators, the spectral gap of the 2-ASEP in the thermodynamic limit, exhaustive stationary phase diagrams, and the extension of the reflection-equation framework to the unequal-rate 2-TASEP and to the bridge model with spontaneous symmetry breaking. Later literature settled the two-species stationary states, produced N-species K-matrix classes and Bethe equations for two of the four boundary classes per side; the remaining core remains open.

Background

The asymmetric simple exclusion process (ASEP) is the standard stochastic model of biased one-dimensional transport: particles hop to neighboring lattice sites with at most one particle per site (the exclusion constraint), with rate 1 to the right and rate $q<1$ to the left; the limit $q\to 0$ is the totally asymmetric TASEP. An open chain, coupled at its two ends to particle reservoirs, is the canonical boundary-driven nonequilibrium system: its exact solution by the matrix product ansatz (stationary probabilities written as traces of products of non-commuting matrices obeying the quadratic algebra of Derrida, Evans, Hakim and Pasquier (DEHP)) and by Bethe ansatz (the spectrum of the Markov generator, including its spectral gap, i.e. the smallest nonzero eigenvalue which sets the relaxation time, computed in the thermodynamic limit by de Gier and Essler) is a cornerstone of nonequilibrium statistical mechanics.

The two-species ASEP (2-ASEP) carries three species per site, denoted 3, 2, 1: species 3 is a first-class particle, species 2 a second-class particle (the carrier of shocks, i.e. of moving interfaces between domains of different density), and species 1 plays the role of a hole. Two neighboring particles are exchanged with rate 1 if they appear in decreasing order and with rate $q$ otherwise. Crampe, Mallick, Ragoucy and Vanicat (2015) asked which boundary injection/extraction rates of the open 2-ASEP are integrable, meaning compatible with the Yang-Baxter equation of the bulk $R$-matrix (the matrix encoding the local transition rates) and with Sklyanin's reflection equation for boundary $K$-matrices, whose consequence is a commuting family of transfer matrices containing the Markov generator. Imposing that the $K$-matrices be unitary and Markovian (generating honest stochastic boundary rules), they classified the solutions completely: four boundary classes on each side, hence sixteen integrable 2-TASEP models, some with all particle currents non-vanishing so that the models are genuinely out of equilibrium. For one representative model they constructed the stationary state through a matrix ansatz built from two algebras familiar from integrable quantum field theory: the Zamolodchikov-Faddeev (ZF) algebra, giving the bulk exchange relations among nine generators, and the Ghoshal-Zamolodchikov (GZ) algebra, giving the action of the generators on two boundary vectors; all stationary weights follow from these relations.

The conclusion of that paper posed five linked open questions that define the problem stated here: (i) an exhaustive phenomenological analysis of all boundary types of the classification; (ii) explicit concrete representations of the ZF/GZ algebras, plausibly as tensor products of the $q$-deformed oscillator algebras used for the periodic $N$-species problem; (iii) the spectral gap of the 2-ASEP, known for the single-species ASEP; (iv) integrable boundaries for generalizations, namely the 2-TASEP with unequal species-dependent hopping rates, the homogeneous $N$-species ASEP for general $N$ (all species hopping with the same rates), and the bridge model, a 2-TASEP variant whose boundary rates can induce spontaneous symmetry breaking and which has eluded exact solution; and (v) the generalization of (modified) algebraic Bethe ansatz techniques to diagonalize the transfer matrices of these multi-species models with generic integrable boundaries.

Subsequent work settled substantial parts of this programme: two-species matrix product stationary states via tensor products of DEHP algebras; classes of $N$-species integrable $K$-matrices embedded in an algebra generalizing the boundary Hecke algebra (the quotient of the type-A Hecke algebra that encodes boundary conditions in reflection-equation language); a proof that the normalization of the open multi-species ASEP stationary state is a Koornwinder polynomial (a multivariable orthogonal polynomial associated with the type-C root system); nested off-diagonal Bethe ansatz equations (from a Bethe-ansatz scheme designed for transfer matrices with non-diagonal boundary terms) for two of the four boundary classes per side; and, recently, type-C multiline queues (combinatorial objects generalizing the multiline queues introduced by Ferrari and Martin to compute the stationary distribution of the periodic multispecies TASEP) for one boundary specialization. What remains open is stated in the problem below, with the literature lineage in the Background paragraph above.

Problem Statement

The problem is to complete the exact-solution programme for open multi-species exclusion processes with integrable boundaries, i.e. boundary rates belonging to the classes classified by solving Sklyanin's reflection equation with unitary Markovian $K$-matrices. For the 2-ASEP (species 3, 2, 1 = first-class particle, second-class particle, hole; bulk exchange rates 1 and $q$) and its $N$-species generalizations, four linked deliverables are requested, sharing one objective, the exact solvability of boundary-driven multi-species exclusion dynamics:

(a) Stationary states and boundary algebras. Construct explicit representations of the ZF/GZ boundary algebras, equivalently explicit matrix product or combinatorial stationary states with computable normalization, species densities, currents and correlations, for all integrable boundary classes and arbitrary species number, beyond the two-species cases already settled; in particular for the $N$-species integrable classes ($N\ge 3$, generic boundary parameters), including the multi-species ASEP whose stationary normalization is a Koornwinder polynomial.

(b) Spectra and the gap. Diagonalize the Markov generator of the integrable open 2-ASEP by Bethe ansatz for the boundary classes not yet treated (the remaining two of the four classes per side), obtaining eigenvectors and eigenvalues, and determine the spectral gap of the 2-ASEP in the thermodynamic limit as a function of the boundary parameters, generalizing the single-species result; extend the diagonalization to the $N$-species models.

(c) Phenomenology. Derive from the exact stationary states the complete phase diagrams (phases, coexistence lines, shock structure as functions of the boundary parameters) for the full boundary classification, replacing the current case-by-case coverage.

(d) Variant models. Extend the reflection-equation classification and the matrix ansatz to the signposted generalizations: the 2-TASEP with unequal species-dependent hopping rates (integrable bulk dynamics, boundaries unknown) and the bridge model whose boundary-induced spontaneous symmetry breaking has eluded exact solution; for this part, a rigorous obstruction result (non-existence of Markovian integrable boundaries for these bulk dynamics, or non-integrability of the bridge model) would also constitute a resolution.

These deliverables are jointly the field's standard meaning of an exact solution, stationary state, spectrum and phase structure, obtained through a single reflection-equation framework: (a) is the algebraic core, (b) its spectral counterpart, (c) its physical consequence, and (d) the extension of the same machinery to variant bulk rules. Completion is judged per model family: for each integrable boundary class and species number, the family is solved when its stationary state, its spectrum (or at least its spectral gap), and its phase structure are known exactly. Full resolution requires all four deliverables; a model family solved under this rule counts as partial progress, not as resolution of the task. For the variant models of (d) the corresponding criterion is completion of their boundary classification, or a rigorous obstruction to it. Components already settled in the literature (two-species stationary states, the $N$-species $K$-matrix classification, Bethe equations for two of the four boundary classes per side) are recorded in the Background above and excluded from the target. The Koornwinder-polynomial normalization of the open multi-species ASEP and the type-C multiline-queue construction (which covers one boundary specialization only), also named in the Background, are excluded in the same way: for those models deliverable (a) still requires the stationary state itself in matrix product or combinatorial form, not its normalization, and not only at the boundary specialization already covered.

References

1. N. Crampe, K. Mallick, E. Ragoucy, M. Vanicat, "Open two-species exclusion processes with integrable boundaries", J. Phys. A: Math. Theor. 48 (2015) 175002. DOI: 10.1088/1751-8113/48/17/175002, arXiv:1412.5939, https://arxiv.org/abs/1412.5939 (source of the open programme)
2. N. Crampe, C. Finn, E. Ragoucy, M. Vanicat, "Integrable boundary conditions for multi-species ASEP", J. Phys. A: Math. Theor. 49 (2016) 375201. DOI: 10.1088/1751-8113/49/37/375201, arXiv:1606.01018, https://arxiv.org/abs/1606.01018
3. N. Crampe, M. R. Evans, K. Mallick, E. Ragoucy, M. Vanicat, "Matrix product solution to a 2-species TASEP with open integrable boundaries", J. Phys. A: Math. Theor. 49 (2016) 475001. DOI: 10.1088/1751-8113/49/47/475001, arXiv:1606.08148, https://arxiv.org/abs/1606.08148
4. C. Finn, E. Ragoucy, M. Vanicat, "Matrix product solution to multi-species ASEP with open boundaries", J. Stat. Mech. (2018) 043201. DOI: 10.1088/1742-5468/aab1b5, arXiv:1712.06809, https://arxiv.org/abs/1712.06809
5. L. Cantini, A. Garbali, J. de Gier, M. Wheeler, "Koornwinder polynomials and the stationary multi-species asymmetric exclusion process with open boundaries", J. Phys. A: Math. Theor. 49 (2016) 444002. DOI: 10.1088/1751-8113/49/44/444002, arXiv:1607.00039, https://arxiv.org/abs/1607.00039
6. X. Zhang, F. Wen, J. de Gier, "T-Q relations for the integrable two-species asymmetric simple exclusion process with open boundaries", J. Stat. Mech. (2019) 014001. DOI: 10.1088/1742-5468/aaeb4a, arXiv:1808.05880, https://arxiv.org/abs/1808.05880
7. N. Crampe, E. Ragoucy, V. Rittenberg, M. Vanicat, "Integrable dissipative exclusion process: Correlation functions and physical properties", Phys. Rev. E 94 (2016) 032102. DOI: 10.1103/PhysRevE.94.032102, arXiv:1603.06796, https://arxiv.org/abs/1603.06796
8. M. Vanicat, "Exact solution to integrable open multi-species SSEP and macroscopic fluctuation theory", J. Stat. Phys. 166 (2017) 1129-1150. DOI: 10.1007/s10955-016-1705-7, arXiv:1610.08388, https://arxiv.org/abs/1610.08388
9. L. Cantini, "Open 2-TASEP with integrable boundaries", J. Phys. A: Math. Theor. 57 (2024) 145201. DOI: 10.1088/1751-8121/ad2cb2, https://doi.org/10.1088/1751-8121/ad2cb2
10. O. Mandelshtam, J. Valencia-Porras, "Type $C$ multiline queues and the open-boundary TASEP", arXiv:2607.24637 (2026), https://arxiv.org/abs/2607.24637
11. G. M. Schütz, A. Zahra, "Hydrodynamics and boundary-induced phase transitions in the $n$-species particle-exchange process", arXiv:2605.07615 (2026), https://arxiv.org/abs/2605.07615
12. A. Ayyer, D. Roy, "The exact phase diagram for a class of open multispecies asymmetric exclusion processes", Sci. Rep. 7 (2017) 13555. DOI: 10.1038/s41598-017-12768-8, arXiv:1611.01943, https://arxiv.org/abs/1611.01943
13. J. de Gier, F. H. L. Essler, "Slowest relaxation mode of the partially asymmetric exclusion process with open boundaries", J. Phys. A: Math. Theor. 41 (2008) 485002. DOI: 10.1088/1751-8113/41/48/485002, https://doi.org/10.1088/1751-8113/41/48/485002
14. D. Schmid, Z. Yang, "The open ASEP with light particles", Probab. Theory Relat. Fields 195 (2026) 1455-1503. DOI: 10.1007/s00440-026-01486-w, arXiv:2409.00849, https://arxiv.org/abs/2409.00849