Stochastic gyration driven by dichotomous noises

Basic Information

• Paper ID: 2510.13589
• Title: Stochastic gyration driven by dichotomous noises
• Authors: Timothée Herbeau, Leonid Pastur, Pascal Viot, Gleb Oshanin
• Classification: cond-mat.stat-mech (Condensed Matter Physics - Statistical Mechanics)
• Publication Date: October 15, 2025
• Paper Link: https://arxiv.org/abs/2510.13589v1

Abstract

This paper investigates the stochastic dynamics of a planar particle driven by two dichotomous noises in a parabolic confining potential, which serves as an analog to the experimentally relevant Brownian gyrator (BG) model. Unlike the standard BG model, this work assumes that the temporal evolution of position components is driven not by Gaussian white noise, but by two statistically independent dichotomous noises. The authors analytically compute position variances and cross-correlations, as well as mean angular momentum, thereby determining the conditions for spontaneous rotational motion of the particle around the origin. Numerical analysis of mean angular velocity is also performed. Finally, analytical calculations of marginal position probability density functions are provided, revealing extremely rich behavior emerging in this system of two coupled linear stochastic differential equations. The study demonstrates that depending on parameter values characterizing the noise, these distributions converge to stationary forms defined on finite support with highly unusual shapes, featuring multiple maxima and minima, plateau regions, and exhibiting discontinuous behavior.

Research Background and Motivation

Importance of the Problem

1. Universality of Non-equilibrium Systems: Non-equilibrium systems have important applications in biological molecular motors, active cytoskeletal transport, synthetic nanomachines, and other fields. These systems can maintain currents, perform mechanical work, and maintain steady states far from thermodynamic equilibrium.
2. Limitations of the Standard Brownian Gyrator: Although the standard BG model is simple, it has important deficiencies. Under Gaussian white noise driving, the probability density function of angular velocity exhibits heavy tails P(W) ∼ 1/|W|³, meaning only the first moment exists. Fluctuations completely destroy the system's rotational motion, preventing the BG from functioning as a conventional heat engine.
3. Experimental Feasibility: In optical trapping of particles or equivalent circuit experiments, noise is computer-generated, so in principle one can design noise with arbitrary statistical properties, providing an experimental foundation for studying dichotomous noise-driven systems.

Research Motivation

The authors aim to "tune" the standard BG model by replacing Gaussian white noise with dichotomous noise, seeking parameter ranges that make the particle position probability density function have a minimum value at the origin (ideally zero), thereby expecting to obtain more regular angular velocity behavior and more systematic rotational motion.

Core Contributions

1. Established a new Stochastic Gyrator (SG) model: Proposed a mathematical framework of two coupled generalized Ornstein-Uhlenbeck processes driven by dichotomous noise.
2. Analytically computed key statistical quantities: Obtained exact expressions for second moments of position components, cross-moments, and mean angular momentum.
3. Determined conditions for spontaneous rotation: Established parameter conditions for the system to generate non-zero mean angular momentum and angular velocity.
4. Revealed rich probability distribution behavior: Discovered six distinct parameter regions, each corresponding to unique probability density function shapes, including multimodal, plateau, and discontinuous complex behaviors.
5. Provided possibility of avoiding heavy-tail distributions: In certain parameter regions, the probability of the particle at the origin is zero, potentially eliminating the heavy-tail characteristics of angular velocity distribution in the standard BG model.

Methodology Details

Problem Formulation

Investigation of particle motion in a parabolic confining potential U = (x² + y²)/2 - uxy, whose dynamics are described by the coupled stochastic differential equations:

ẋ(t) = -x(t) + uy(t) + ζₓ(t), x(0) = 0
ẏ(t) = -y(t) + ux(t) + ζᵧ(t), y(0) = 0

where |u| < 1 is the coupling parameter, and ζₓ(t) and ζᵧ(t) are statistically independent symmetric dichotomous noises.

Dichotomous Noise Model

Dichotomous noise is defined as:

ζₐ(t) = {+vₐ, probability = 1/2
         {-vₐ, probability = 1/2

Its autocorrelation function is:

⟨ζₐ(t)ζᵦ(t')⟩ = δₐ,ᵦv²ₐexp(-2λₐ|t-t'|)

where λₐ is the switching rate and τₐ = 1/(2λₐ) is the finite correlation time.

Model Architecture

1. Trajectory Solutions: By solving the coupled equations using standard methods:

   x(t) = ∫₀ᵗ Qc(t-τ)ζₓ(τ)dτ + ∫₀ᵗ Qs(t-τ)ζᵧ(τ)dτ
   y(t) = ∫₀ᵗ Qs(t-τ)ζₓ(τ)dτ + ∫₀ᵗ Qc(t-τ)ζᵧ(τ)dτ
   

   
   where Qc(t) = e⁻ᵗcosh(ut), Qs(t) = e⁻ᵗsinh(ut).
2. Variable Transformation: Introducing S = x + y and D = x - y, simplifying to:

   Ṡ(t) = -(1-u)S(t) + (ζₓ(t) + ζᵧ(t))
   Ḋ(t) = -(1+u)D(t) + (ζₓ(t) - ζᵧ(t))
   

Technical Innovations

1. Bounded Trajectories: Unlike the standard BG model, the SG position components evolve within bounded regions, with region size growing with time and approaching finite limits.
2. Parameter Classification: Introducing αₓ = λₓ/(1-u), αᵧ = λᵧ/(1-u), behavior is classified into six cases (A-F) based on the relationship of these parameters to unity.
3. Noise-Position Correlation: The instantaneous values of dichotomous noise exhibit statistical correlation with position components, which does not exist in the standard BG model.

Experimental Setup

Numerical Simulation Methods

• Numerical solution of stochastic differential equation systems
• Generation of sufficiently long trajectories for statistical computation
• Averaging over 100 trajectory realizations

Analytical Verification

• Verification of analytical expressions for probability density functions through numerical integration
• Complex integral calculations using Mathematica and Scipy

Parameter Settings

Multiple representative parameter sets were used in the paper:

• Coupling parameter: u = 1/2, 3/4
• Noise amplitudes: vₓ, vᵧ ∈ 1, 3
• Switching rates: λₓ, λᵧ ∈ 0.2, 10

Experimental Results

Main Results

1. Angular Momentum and Angular Velocity

In the t → ∞ limit, the mean angular momentum is:

⟨L⟩ = u(Dₓλ²ₓ/((1+λₓ)²-u²) - Dᵧλ²ᵧ/((1+λᵧ)²-u²))

The condition for non-zero mean angular momentum is found to be: u ≠ 0 and (λₓ ≠ λᵧ or vₓ ≠ vᵧ).

2. Six Cases of Probability Density Functions

According to the values of αₓ and αᵧ, P(S) exhibits distinctly different behaviors:

Case A (αₓ < 1, αᵧ < 1):

• Multimodal distribution with divergences at irregular points and support boundaries
• For example, bimodal structure when αₓ = αᵧ = 1/2

Case B (αₓ < 1, αᵧ = 1):

• Combination of plateau and U-shaped structures
• Sharp peaks appear at S = ±Ω

Case C (αₓ < 1, αᵧ > 1):

• Bell-shaped or M-shaped distributions depending on relative magnitudes of vₓ and vᵧ

Case D (αₓ = αᵧ = 1):

• Trapezoidal distribution, degenerating to triangular when vₓ = vᵧ

Case E (αₓ = 1, αᵧ > 1):

• Similar to Case D but with rounded corners

Case F (αₓ > 1, αᵧ > 1):

• Standard bell-shaped distribution, approaching Gaussian form

3. Special Case Analysis

• λₓ = 0 case: Discovery of finite intervals where P(S) ≡ 0, potentially eliminating heavy tails in angular velocity distribution
• Mixed cases: Complex behavior when one noise is Gaussian white noise and the other is dichotomous

Ablation Studies

The paper validates results through:

1. Recovery of standard BG model results in the high switching rate limit
2. Recovery of independent Ornstein-Uhlenbeck processes when u = 0
3. Good agreement between numerical simulations and analytical results

Related Work

Brownian Gyrator Research

• Original Model: First introduced by Exartier and Peliti (1999), developed by Filliger and Reimann (2007)
• Experimental Implementation: Circuit implementation by Ciliberto et al., optical trapping experiments by Argun et al.
• Theoretical Development: Including quantum versions, inertial effects, anisotropic fluctuations, and other extensions

Dichotomous Noise Research

• Classical Theory: Pioneering work by van Kampen, Hänggi, Jung, and others
• Modern Applications: Resurgence in run-and-tumble dynamics and active systems
• Mathematical Methods: Exact results by Sancho (1984) on dichotomous noise-driven processes

Conclusions and Discussion

Main Conclusions

1. Successfully constructed the SG model: The dichotomous noise-driven system exhibits richer dynamical behavior than the standard BG model
2. Discovered diverse probability distributions: Six distinct parameter regions correspond to unique distribution shapes and properties
3. Provided possibilities for improving heat engine performance: In certain parameter regions where the probability at the origin is zero, more systematic rotational motion may be achieved

Limitations

1. Complexity of theoretical analysis: Complete analytical calculations of joint probability density functions and angular velocity distributions remain difficult
2. Experimental verification: Although experimentally feasible in principle, experimental verification with precise control of dichotomous noise parameters is needed
3. Parameter tuning: Precise parameter selection is required to achieve desired dynamical behavior

Future Directions

1. Complete probability distributions: Compute complete position probability density functions P(x,y) and angular velocity distribution P(W)
2. Experimental implementation: Realize dichotomous noise-driven SG in optical trapping or circuit systems
3. Multi-particle systems: Study interactions and synchronization phenomena among multiple SG particles

In-Depth Evaluation

Strengths

1. Strong methodological innovation: First systematic study of dichotomous noise-driven gyrator systems, discovering rich new phenomena
2. Rigorous mathematical treatment: Provides detailed analytical calculations and complete mathematical derivations
3. Comprehensive and in-depth results: Encompasses comprehensive analysis from basic statistical quantities to complex probability distributions
4. Clear physical picture: Effectively explains physical behavior corresponding to different parameter regions

Weaknesses

1. Lack of experimental verification: Although experimental feasibility is mentioned, actual experimental verification is absent
2. Limited application scenarios: Primarily focused on theoretical analysis with limited discussion of practical applications
3. Computational complexity: In certain cases numerical integration is required, limiting the universality of analytical results

Impact

1. Theoretical contribution: Provides new models and analytical methods for non-equilibrium statistical mechanics
2. Methodological value: Dichotomous noise handling techniques can be applied to other stochastic systems
3. Experimental guidance: Provides theoretical foundation for designing microscopic heat engines with better performance

Applicable Scenarios

1. Microscopic heat engine design: Provides guidance for developing nanoscale heat engines with more stable rotational performance
2. Active matter research: Can be applied to studying active systems such as run-and-tumble particles
3. Stochastic process theory: Provides examples for studying stochastic systems driven by non-Gaussian noise

References

The paper includes 72 references covering important works in non-equilibrium statistical mechanics, stochastic thermodynamics, Brownian motors, dichotomous noise theory, and related fields, providing a solid theoretical foundation for the research.