One-Dimensional Stirring Process

• The one-dimensional stirring process is a stochastic model on discrete lattices that combines local particle exchanges with boundary events like creation and annihilation.
• It employs diffusive rescaling to rigorously derive macroscopic transport equations and nonlinear boundary conditions, illuminating nonequilibrium behavior.
• Analytical tools such as duality techniques and SPDE limits yield explicit expressions for correlations, fluctuations, and invariant measures in these systems.

The one-dimensional stirring process encompasses a class of stochastic particle systems on discrete lattices in which local random exchanges (stirring) are combined with non-conservative events such as creation, annihilation, or reservoir-driven particle flux. These models serve as canonical nonequilibrium systems, bridging statistical mechanics, probability, and hydrodynamic theory, and are central in the derivation of macroscopic transport equations, fluctuation theory, and the study of long-range correlations in driven diffusive media.

1. Model Definitions and Fundamental Dynamics

The prototypical one-dimensional stirring process is defined on a finite lattice segment, commonly $\Lambda_N = \{-N, ..., N\} \subset \mathbb{Z}$, with configurations $n \in \{0,1\}^{\Lambda_N}$ representing at most one particle per site. Dynamics are generated by a Markov operator comprising:

• Bulk stirring (exclusion dynamics): Particles on neighboring sites $(x, x+1)$ exchange position at a rate of $1/2$, enforced by the generator

$$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$

where $n^{x,x+1}$ denotes the configuration after swapping occupations at $x$ and $x+1$.

• Boundary birth and death mechanisms: At designated intervals near boundaries, $I_+ = (N-K, N]$ (births) and $I_- = [-N, -N+K)$ (deaths), particles are injected or removed via mechanisms acting on the rightmost empty (for births) or leftmost occupied (for deaths) site within the respective interval. For example, the birth part reads

$n \in \{0,1\}^{\Lambda_N}$0

where $n \in \{0,1\}^{\Lambda_N}$1 enforces constraint on which site in $n \in \{0,1\}^{\Lambda_N}$2 is eligible for a birth event (Masi et al., 2011, Birmpa et al., 15 Dec 2025).

Time-rescaling by $n \in \{0,1\}^{\Lambda_N}$3 leads to the full generator

$n \in \{0,1\}^{\Lambda_N}$4

which sets the stirring to be the dominant process on diffusive time scales, aligning the microscopic dynamics with macroscopic transport behavior.

The process can be generalized to multispecies settings, higher dimensions, or to incorporate alternative boundary conditions, such as open chains with reservoirs or periodic rings (Casini et al., 2023, Ioffe et al., 2019).

2. Hydrodynamic Limit and Macroscopic Equations

In the hydrodynamic scaling limit ($n \in \{0,1\}^{\Lambda_N}$5, $n \in \{0,1\}^{\Lambda_N}$6, diffusive rescaling), empirical measures of particle density converge, in probability, to deterministic macroscopic profiles governed by reaction-diffusion or heat equations with nontrivial boundary conditions:

• Hydrodynamic PDE: For the single-species stirring process with boundary births and deaths,

$n \in \{0,1\}^{\Lambda_N}$7

on $n \in \{0,1\}^{\Lambda_N}$8, supplemented by nonlinear Robin-type boundary data:

$n \in \{0,1\}^{\Lambda_N}$9

The initial profile $(x, x+1)$0 is inherited from the initial microscopic measure (Birmpa et al., 15 Dec 2025).

This scaling framework provides the rigorous derivation of macroscopic transport laws from microscopic stochastic rules, enabling the identification of effective boundary conditions and revealing the nonlinear response induced by creation-annihilation mechanisms at the boundaries (Masi et al., 2011, Birmpa et al., 15 Dec 2025).

3. Correlation Hierarchies and Factorization Properties

Microscopic occupation variables develop correlations due to stirring and boundary interactions. Central to the analysis is the study of truncated correlations, or $(x, x+1)$1-functions,

$(x, x+1)$2

where $(x, x+1)$3 is the local macroscopic profile and $(x, x+1)$4 distinct sites.

Rigorous moment bounds establish:

• For any fixed $(x, x+1)$5, there exists $(x, x+1)$6 such that

$(x, x+1)$7

for $(x, x+1)$8, uniformly in $(x, x+1)$9 (Masi et al., 2011). For large $1/2$0, $1/2$1.

• In the fluctuation regime, for centered occupations $1/2$2, spatial $1/2$3-functions of degree $1/2$4 satisfy

$1/2$5

for some $1/2$6 and all $1/2$7 uniformly (Birmpa et al., 15 Dec 2025).

These bounds imply that, for small $1/2$8, occupations at finitely many distinct sites become asymptotically independent ("propagation of chaos"), permitting strong factorization of finite-dimensional marginals and facilitating Gaussian central limit theorems for empirical sums (Masi et al., 2011, Birmpa et al., 15 Dec 2025).

4. Duality and Exact Solutions for Open Boundary Conditions

Duality techniques, particularly absorption duality, structurally underpin the derivation of closed expressions for stationary states and correlation functions in boundary-driven stirring processes:

• The construction of a dual process on an extended graph with absorbing sites enables the exact representation of non-equilibrium steady-state $1/2$9-point occupation probabilities as weighted absorption probabilities of $$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$0 dual particles (Casini et al., 2023).
• In multispecies hard-core exclusion models, integrability based on underlying $$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$1 symmetry and the Matrix Product Ansatz (MPA) yields explicit closed-form expressions for all correlations:

$$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$2

with all $$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$3-point correlations polynomial in the boundary reservoir fugacities (Casini et al., 2023).

For the single-species symmetric exclusion process, the duality approach recovers the classical linear density profiles and quantifies boundary-induced long-range spatial correlations in the nonequilibrium stationary state.

5. Fluctuations, SPDE Limits, and Ornstein-Uhlenbeck Processes

Starting from product measures associated with smooth profiles, rescaled fluctuations of the density converge to infinite-dimensional Ornstein-Uhlenbeck (OU) processes:

• The fluctuation field, for test function $$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$4, is given by

$$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$5

converging in distribution as $$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$6 to a mean-zero Gaussian process

$$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$7

where $$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$8 is the adjoint semigroup of the linearized heat equation with time-dependent Robin boundary conditions (Birmpa et al., 15 Dec 2025).

The limiting SPDE reads

$$L_0 f(n) = \frac{1}{2} \sum_{x=-N}^{N-1} \left[f(n^{x,x+1}) - f(n)\right],$$9

with $n^{x,x+1}$0 and $n^{x,x+1}$1 restricted to test functions satisfying boundary-linearized constraints. This macroscopic fluctuation theory is underpinned by the refined $n^{x,x+1}$2-function bounds, which ensure vanishing contributions of non-Gaussian higher-order terms.

6. Cycle-Structure, Permutation Dynamics, and Split-Merge Limits

On periodic geometries (discrete ring $n^{x,x+1}$3), the stirring process extends to random interchange models where the configuration space is the permutation group $n^{x,x+1}$4:

• Each edge undergoes transpositions at rate one; cycle-length statistics of permutation evolve under stirring.
• Under appropriate time-rescaling, the empirical cycle-length distribution converges (in Skorokhod topology) to the canonical continuous-time split-and-merge process with generator

$n^{x,x+1}$5

where $n^{x,x+1}$6 merges and $n^{x,x+1}$7 splits mass blocks (Ioffe et al., 2019).

• The unique invariant measure is the Poisson-Dirichlet law $n^{x,x+1}$8 with explicit stick-breaking and Poisson-point constructions. The stationary cycle-count for $n^{x,x+1}$9-cycles is asymptotically Poisson with mean $x$0.

These results generalize across all $x$1, with intricate combinatorial and coupling arguments required for $x$2 due to recurrence and local interaction constraints.

7. Mechanisms in Quantum Circuits and Atom Flow via Stirring

In quantum atomtronic circuits, such as Bose-Einstein condensates confined to "racetrack" or ring geometries, one-dimensional effective stirring produces quantized irreversible flows:

• The dynamics are governed by the one-dimensional Gross–Pitaevskii equation with a time-dependent potential barrier acting as the stirrer (Eller et al., 2020).
• Macroscopic flow (quantized circulation) is generated by a deterministic sequence of vortex/antivortex nucleation events ("phase slips") triggered when the moving barrier exceeds a critical strength:

$x$3

where $x$4 is stirrer speed and $x$5 is the sound speed.

• The winding number $x$6 of phase slips is set by

$x$7

leading to final mean flow $x$8.

• Thermal effects reduce the critical barrier, induce earlier nucleation, and smooth flow, but quantized steps in circulation persist up to substantial fractions of $x$9.

These mechanisms establish a controlled pathway for engineering persistent flows in one-dimensional atom circuits via tailored stirring protocols (Eller et al., 2020).

---

References:

• (Masi et al., 2011) Truncated correlations in the stirring process with births and deaths
• (Birmpa et al., 15 Dec 2025) Non-equilibrium fluctuations for the stirring process with births and deaths
• (Casini et al., 2023) Duality for the multispecies stirring process with open boundaries
• (Ioffe et al., 2019) Split-and-Merge in Stationary Random Stirring on Lattice Torus
• (Eller et al., 2020) Producing flow in "racetrack" atom circuits by stirring at zero and non-zero temperature