Hydrodynamics of Multi-Species Driven Diffusive Systems with Open Boundaries: A Two-Tasep Study

In this short note, we review a recently developed method for analysing multi-component driven diffusive systems with open boundaries. The approach generalises the extremal-current principle known for single-component models and is based on solving the Riemann problem for the corresponding hydrodynamic equations. As a case study, we focus on a two-species exclusion process on a lattice (Two-TASEP), where two types of particles move in opposite directions with two arbitrary rates and exchange positions upon encounter with a third rate. Despite its simplicity, this toy model effectively captures the key features of multi-species driven diffusive systems, including phase separation phenomena. This allows us to illustrate the critical role played by the underlying Riemann invariants in determining the system’s macroscopic behavior.

💡 Research Summary

The paper presents a unified hydrodynamic framework for analyzing multi‑species driven diffusive systems with open boundaries, extending the well‑known extremal‑current principle from single‑species models to systems containing several interacting particle types. The central claim is that the stationary bulk densities of an open system are given by the solution at the origin of the corresponding Riemann problem for the conservation‑law system ∂t ρ + ∂x J(ρ)=0, where the left and right initial states are the densities imposed by the two reservoirs. For a single‑species totally asymmetric simple exclusion process (TASEP) this statement reduces to the familiar extremal‑current rule, but the authors argue that it holds in full generality for any number of conserved species.

To illustrate the method, the authors study a two‑species exclusion process (Two‑TASEP). In this model particles of type “•” hop rightward with rate β, particles of type “◦” hop leftward with rate α, and when a • and a ◦ meet they exchange positions at unit rate (time scale chosen so that the exchange rate is 1). Empty sites are denoted by *. The microscopic dynamics are simple, yet the stationary measure is not of product form except for special parameter choices (α=β=½). Consequently, the macroscopic currents cannot be written as a mean‑field function of the densities.

Using the nested algebraic Bethe Ansatz, the authors obtain implicit expressions for the currents: J◦ = zα (zβ−1) + ρ◦ (zα−zβ), J• = zβ (1−zα) + ρ• (zα−zβ), where the auxiliary variables zα, zβ lie in a bounded domain and are linked to the physical densities through two nonlinear equations (5)–(6). Remarkably, the level sets of constant zα (or constant zβ) are straight lines in the (ρ•, ρ◦) plane. These lines satisfy the Rankine‑Hugoniot condition and therefore constitute shock curves; because the system belongs to the Leroux class, the same lines also serve as rarefaction curves. Hence the model possesses two Riemann invariants, namely zα and zβ.

By differentiating the implicit relations, the authors derive decoupled conservation laws for the invariants: ∂t zα + vα(z) ∂x zα = 0, ∂t zβ + vβ(z) ∂x zβ = 0, with characteristic speeds vα, vβ given by the eigenvalues of the Jacobian ∂ρ J. The eigenvectors are orthogonal to the gradients of the opposite invariant, confirming that (zα, zβ) are indeed the Riemann variables. This diagonalisation shows that the system is of “T‑temple” type: shock and rarefaction curves coincide and are linear.

Armed with the Riemann invariants, the authors address the open‑boundary problem. The left and right reservoirs impose effective densities ρL and ρR, which can be mapped to corresponding invariant values (zαL, zβL) and (zαR, zβR). Solving the Riemann problem with these step initial conditions yields a self‑similar solution ρ(x/t). Evaluating this solution at x=0 provides the bulk densities (ρ•, ρ◦) that the open system will select. Depending on the sign of the characteristic velocities at the bulk state, three macroscopic phases arise:

1. Left‑Induced (LI) phase: bulk density equals the left reservoir density, characteristic speed vB>0, perturbations propagate from left to bulk.
2. Right‑Induced (RI) phase: bulk density equals the right reservoir density, vB<0, perturbations propagate from right.
3. Bulk‑Induced (BI) phase: bulk density is independent of both reservoirs, vB=0; this corresponds to a maximal‑current‑like state where the invariant values satisfy zα=½ or zβ=½.

These phases generalise the low‑density, high‑density, and maximal‑current phases of the single‑species TASEP, but now each phase is characterised by the behaviour of both invariants. The authors also discuss how the shock admissibility conditions (ordering of invariant values across a shock) select physically relevant solutions.

The paper concludes that the Riemann‑invariant approach provides a systematic, non‑perturbative method to predict phase diagrams, bulk currents, and density profiles for multi‑species driven diffusive systems with arbitrary bulk and boundary rates, without requiring integrable boundaries or product‑measure stationary states. The authors suggest that this framework can be directly applied to biologically relevant transport processes, such as multiple molecular motor species moving along microtubules or axonal transport, where different particle types have preferred directions and interact upon encounter.