Multistability and Control in Ring Networks of Phase Oscillators with Frequency Heterogeneity and Phase Lag

This paper numerically investigates a ring network of phase oscillators with frequency heterogeneity and phase lag, revealing how these factors influence the basin sizes of multistable synchronous states and proposing a control method to steer the system toward specific wavenumbers.

The Gist

Imagine a large circle of dancers, each holding a flashlight. In a perfect world, they all spin at the exact same speed and flash their lights in perfect unison. This is synchronization. But in the real world, things are messy. Some dancers are naturally faster, some are slower (this is frequency heterogeneity), and sometimes they react to their neighbor's move a split second too late or too early (this is phase lag).

This paper explores what happens when these messy dancers try to coordinate. The surprising discovery? The group doesn't just have one "correct" way to dance. It has many.

The Many Ways to Dance: "Twisted" States

In a perfect circle, the dancers could all flash together (0 twists). But they could also flash in a wave that travels around the circle once (1 twist), twice (2 twists), or even in the opposite direction.

Think of these like different spirals in a galaxy or different knits in a sweater.

• Low Wavenumber (0 or 1): The pattern is simple, like a flat circle or a gentle wave.
• High Wavenumber (5 or 6): The pattern is complex, with many tight twists and turns.

The paper calls this Multistability. It means the system is like a ball sitting on a landscape with many valleys. Depending on where you start (the initial conditions), the ball will roll into a different valley. Once it's there, it stays there.

The "Valley" Size: Who Wins the Lottery?

The researchers wanted to know: How likely is the group to end up in a specific pattern?

They imagined throwing a dart at a map of all possible starting positions. The size of the "target area" (called a basin) for each pattern tells you how likely that pattern is to happen by chance.

• Big Basin: Easy to fall into. If you start almost anywhere, you end up here.
• Small Basin: Hard to hit. You need to start in a very specific spot to end up here.

The Discovery:

1. The "Lag" Effect: When the dancers have a slight delay in reacting to each other (phase lag), it actually helps the complex, high-twist patterns (the "high wavenumbers") become bigger and more stable. It's like adding a little bit of friction to a spinning top; it surprisingly helps it spin in a more complex way.
2. The "Noise" Effect: When the dancers have different natural speeds (heterogeneity), it usually destroys the complex patterns first. The complex, high-twist spirals are fragile. The simple, flat patterns (low wavenumbers) are tough and survive even when the group is messy.

The Counter-Intuitive Twist

Here is the weird part: The paper found that adding a little bit of messiness (different speeds) actually makes the simple patterns even more dominant. It's like saying, "If everyone is a bit out of sync, the group is more likely to settle into a simple, flat circle than a complex spiral." The complex spirals are too fragile to survive the chaos.

The Magic Control Switch

The most exciting part is the solution. The researchers figured out how to steer the dancers into a specific pattern without touching them individually.

Imagine you want the group to form a specific 3-twist spiral, but they keep falling into a 0-twist flat circle.

1. The Trick: You temporarily introduce a specific amount of "messiness" (heterogeneity) and "delay" (phase lag).
2. The Result: This specific combination of messiness acts like a filter. It destroys the "flat circle" valley and the "5-twist" valley, leaving only the "3-twist" valley standing.
3. The Switch: The group is forced to roll into the 3-twist pattern.
4. The Cleanup: Once they are locked into the 3-twist pattern, you remove the messiness. Because the pattern is now stable, they stay there!

Why Does This Matter?

This isn't just about dancers. It applies to:

• The Heart: Sometimes heart cells beat in a chaotic, dangerous pattern (fibrillation). This research helps us understand how to nudge the heart back into a healthy rhythm.
• Power Grids: Ensuring all the generators in a city stay in sync so the lights don't go out.
• Robot Swarms: Getting a group of robots to move in a specific formation without a central commander telling every single robot what to do.

In short: The paper teaches us that by carefully tuning the "noise" and "delay" in a system, we can force a chaotic group of oscillators to pick the exact pattern we want, turning a multi-option mess into a single, controlled solution.

Technical Summary

Here is a detailed technical summary of the paper "Multistability and Control in Ring Networks of Phase Oscillators with Frequency Heterogeneity and Phase Lag" by Soomin Kim and Hiroshi Kori.

1. Problem Statement

Oscillator networks, such as neural circuits, power grids, and biological rhythms, often exhibit multistability, where multiple distinct synchronized states (attractors) coexist. The system's final state depends heavily on initial conditions.

• The Challenge: In practical applications (e.g., cardiac fibrillation vs. normal rhythm, or switching gaits in locomotion), it is crucial to understand the accessibility of these states (how likely a specific pattern is to emerge) and to develop methods to steer the system toward a desired state without direct manipulation of individual oscillator phases.
• The Gap: Previous studies (e.g., Wiley and Strogatz) analyzed multistability in homogeneous ring networks. However, real-world systems inherently possess frequency heterogeneity (natural frequency variations) and often involve phase lags in interactions. The impact of these two factors on the basin of attraction (the set of initial conditions leading to a specific state) and the robustness of high-wavenumber states was not fully understood.

2. Methodology

The authors extended the classical Kuramoto model for a ring network of $N$ phase oscillators to include two critical realistic features:

1. Frequency Heterogeneity: Natural frequencies $\omega_k$ are drawn from a normal distribution $N(0, \sigma^2)$.
2. Phase Lag: The interaction function includes a constant phase offset $\alpha$ (Sakaguchi–Kuramoto coupling).

The Model:
The dynamics are governed by:
$$\dot{\phi}_k = \omega_k + \sin(\phi_{k-1} - \phi_k + \alpha) + \sin(\phi_{k+1} - \phi_k + \alpha)$$
where $\phi_k$ is the phase of oscillator $k$, and periodic boundary conditions apply.

Numerical Approach:

• System Size: $N = 80$ oscillators.
• Simulation Protocol:
  • Random initial phases and frequency realizations were generated for $M = 10^4$ independent runs.
  • Convergence to a phase-locked state was determined by monitoring the Kuramoto order parameter $r(t)$.
  • The winding number $q$ (representing the spatial wavenumber of the synchronization pattern) was calculated for each converged state.
• Basin Analysis: The "basin size" was quantified as the probability of converging to a specific $q$-twisted state from random initial conditions.
• Robustness Analysis: The critical heterogeneity threshold $\sigma_c(q)$ was determined by incrementally increasing $\sigma$ starting from a stable $q$-twisted state until the state lost phase locking.
• Control Strategy: A parameter-switching protocol was proposed to exploit differences in robustness.

3. Key Contributions

1. Quantification of Basin Size Distribution: The study maps how the distribution of winding numbers changes with phase lag ($\alpha$) and frequency heterogeneity ($\sigma$).
2. Discovery of Counter-Intuitive Robustness: It reveals that while heterogeneity generally destabilizes synchronization, its effect is state-dependent:
   • In homogeneous systems ($\alpha=0$), heterogeneity preferentially destroys high-wavenumber states, making low-wavenumber (near in-phase) states more dominant.
   • With significant phase lag ($\alpha > 0$), the most robust state shifts to non-zero wavenumbers ($|q| > 0$).
3. Parameter-Based Control Protocol: The authors propose a novel control method that does not require actuating individual oscillators. Instead, it temporarily modulates global parameters ($\sigma$ and $\alpha$) to eliminate undesired attractors, forcing the system into a target state.

4. Key Results

A. Effect of Phase Lag ($\alpha$) on Basin Sizes

• Broadening of Distribution: As the phase lag $\alpha$ increases (up to a certain point), the distribution of realized winding numbers $q$ broadens.
• Promotion of High Wavenumbers: Increasing $\alpha$ facilitates convergence to twisted states with higher $|q|$ values. The standard deviation of the $q$ distribution increases monotonically with $\alpha$.
• Optimal Synchronization: The system shows maximum robustness (highest probability of phase-locking) at an intermediate phase lag of $\alpha \approx 0.20\pi$.

B. Effect of Frequency Heterogeneity ($\sigma$)

• State-Dependent Destabilization:
  • Low $\alpha$ (e.g., $\alpha=0$): Increasing $\sigma$ narrows the $q$ distribution, effectively destroying high-wavenumber states and leaving the in-phase state ($q=0$) as the most robust.
  • High $\alpha$ (e.g., $\alpha=0.3\pi$): The relationship becomes non-monotonic. For $\sigma > 0.2$, the distribution actually widens as $\sigma$ increases.
• Critical Thresholds ($\sigma_c$): The critical heterogeneity $\sigma_c(q)$ (the point where a state $q$ loses stability) is unimodal.
  • For $\alpha=0$, $\sigma_c$ peaks at $q=0$.
  • For $\alpha > 0$, the peak of $\sigma_c$ shifts to $|q| > 0$. This implies that phase lag makes twisted states more robust against frequency heterogeneity than the in-phase state.

C. Control Strategy (Parameter Switching)

• Mechanism: The authors exploit the fact that different $q$-states have different tolerances to heterogeneity.
• Protocol:
  1. Evolve the system at baseline parameters ($\sigma=0$).
  2. Temporarily switch to a specific heterogeneity level $\sigma^*$ and phase lag $\alpha$ where only the target state $q^*$ remains stable (all other competing attractors are destabilized).
  3. Once the system converges to $q^*$, switch parameters back to baseline.
• Outcome: Numerical simulations demonstrated that this method successfully steers the system from various random initial conditions to a specific target winding number (e.g., $q=1$) and maintains it after the control signal is removed.

5. Significance

• Theoretical Insight: The paper resolves the complex interplay between non-local phase lags and local frequency disorder in multistable systems. It challenges the intuition that heterogeneity always reduces the basin of complex states; instead, it shows that heterogeneity can selectively prune specific attractors, reshaping the global landscape.
• Practical Application: The proposed control strategy is highly relevant for biological and engineering systems where direct phase control is impossible or energetically costly (e.g., controlling neural rhythms in epilepsy or switching modes in power grids). By tuning global parameters (like coupling strength or noise levels), one can guide the system to a desired functional state.
• Design Principles: The findings suggest that phase lag and frequency heterogeneity can be used as "design knobs" to engineer oscillator networks with specific multistable properties, enhancing the ability to select desired synchronization patterns.

In summary, this work provides a comprehensive framework for understanding and controlling multistability in oscillator networks, demonstrating that the "weakness" of a system (heterogeneity) can be strategically leveraged to achieve robust control.