Abstract
We consider the Czirók model for collective motion of locusts along a one-dimensional torus. In the model, each agent's velocity locally interacts with other agents' velocities in the system, and there is also exogenous randomness to each agent's velocity. The interaction tends to create the alignment of collective motion. By analyzing the associated nonlinear Fokker-Planck equation, we obtain the condition for the existence of stationary order states and the conditions for their linear stability. These conditions depend on the noise level, which should be strong enough, and on the interaction between the agent's velocities, which should be neither too small, nor too strong. We carry out the fluctuation analysis of the interacting system and describe the large deviation principle to calculate the transition probability from one order state to the other. Numerical simulations confirm our analytical findings.
Introduction
Collective motion is a ubiquitous phenomenon in nature, observed in bird flocks, fish schools, and locust swarms. Understanding how these systems maintain order and why they occasionally undergo sudden transitions—such as a swarm reversing its direction—is a central question in biological physics. Traditionally, researchers have used either Self-Propelled Particle (SPP) models, which track every individual, or Coarse-Grained (CG) models, which look at overall density.
However, a "stochastic bridge" has been missing: SPP models are mathematically intractable for large $N$, while standard CG models often average out the noise that drives state transitions (bistability). This paper addresses this gap by applying mean-field theory to the Czirók model, providing a rigorous framework to understand how noise and interaction strength dictate swarm behavior.
Methodology
The study employs the McKean-Vlasov framework to derive a nonlinear Fokker-Planck equation that describes the system as the number of agents ($N$) approaches infinity. The analysis is conducted in three primary stages:
- Linear Stability Analysis: The system is decomposed into Fourier modes to identify the conditions under which uniform collective motion remains stable against spatial perturbations.
- Large Deviation Principle (LDP): To quantify the rare event of a swarm switching directions (e.g., from clockwise to counter-clockwise), the authors use a formal LDP to calculate transition probabilities.
- Numerical Validation: Euler-method discretizations of the underlying stochastic differential equations are used to verify the analytical thresholds and transition frequencies.
The Role of Noise and Interaction
Noise as a Stabilizer
Counter-intuitively, the research reveals that noise acts as a stabilizer for uniform motion. There exists a specific "noise threshold"; if the external noise is too low, the system becomes susceptible to modulational instability. In this regime, agents do not move in a uniform stream but instead bunch up into moving clusters or "bands."
The "Goldilocks" Interaction
Stability is not a monotonic function of interaction strength. The authors find that stationary order states only exist when the interaction between agents falls within a specific middle range. If interactions are too weak, the agents fail to align; if they are too strong, the system becomes unstable to spatial perturbations, leading to clustering.
Key Results and Dynamics
- Stationary States: The authors proved that stationary solutions are uniform in space and Gaussian in velocity. The mean velocity $\xi$ must satisfy a specific compatibility condition, $\xi = G(\xi)$.
- Stability Thresholds: For a potential parameter $h=6$, the calculated noise threshold is $\sigma \approx 0.85$. Simulations confirmed that for $\sigma > 0.85$, the swarm moves uniformly, whereas for $\sigma < 0.85$, the system forms distinct moving bands.
- Exponential Persistence: A critical finding is that the probability of a swarm spontaneously reversing direction decays exponentially with the number of agents ($e^{-N}$). This explains why large biological swarms appear so persistent in their chosen direction despite the inherent randomness of individual members.
- Phase vs. Group Velocity: In the low-noise unstable regime, the "apparent velocity" of a moving cluster (phase velocity) can differ significantly from the average velocity of the individual particles (group velocity).
Conclusion and Future Directions
This work provides a mathematical foundation for the bistability observed in collective motion. It highlights that the persistence of a swarm's direction is a function of its size and the delicate balance between alignment interactions and exogenous noise.
While the current study is limited to a one-dimensional torus and utilizes a formal LDP (as the noise is degenerate, affecting only velocity), it sets the stage for future research. Extending these mean-field results to 2D or 3D spaces and providing a fully rigorous proof for the LDP in the presence of degenerate noise remain important open challenges in the field.
Summary updated