Bellerophon States in Oscillator Networks
- Bellerophon states are structured partially synchronous regimes in coupled oscillators characterized by multi-cluster partitioning and time‐averaged frequency entrainment.
- They are distinct from classical synchronization and chimera states, with clusters showing averaged frequency locking but not full phase or instantaneous frequency alignment.
- Dynamical models like the two-group Kuramoto and Stuart-Landau frameworks illustrate their emergence near criticality, revealing quantized frequency clustering and robust macroscopic order parameters.
Bellerophon states are a class of structured, partially synchronous regimes observed in globally or mean-field coupled oscillator systems with frequency heterogeneity. Distinguished from classical synchronization and chimera states, Bellerophon states exhibit multi-cluster partitioning where oscillators within each cluster are entrained only in their average frequency—neither phases nor instantaneous frequencies lock exactly. This state appears generically near criticality in systems with suitable coupling and frequency distributions, including the Kuramoto model with multi-group structure and the Stuart-Landau network with amplitude degrees of freedom, and persists across various system architectures and order of phase transitions (Bi et al., 2015, Teichmann et al., 2021, Zhang et al., 2019).
1. Defining Characteristics and Distinction from Related States
A Bellerophon state consists of multiple oscillator clusters, each characterized by:
- Shared time-averaged frequency (mean frequency) within clusters, but not shared instantaneous frequency or phase.
- Non-constant instantaneous dynamics; oscillators in a cluster trace out characteristic "cusped" patterns when plotting their instantaneous frequencies against natural frequencies.
- Clusters manifest as discrete "steps" in the long-time average of observed frequency versus natural frequency; between steps, frequency is quasiperiodic.
The Bellerophon state must be distinguished from:
- Classical Kuramoto synchronization: Features full phase or frequency locking as a global, single-cluster regime.
- Chimera states: Require nonlocal or local coupling of identical oscillators and exhibit coexistence of coherent (frequency-locked) and incoherent regions. In contrast, Bellerophon states have global coupling and nonidentical oscillators; all belong to some cluster by time-averaged frequency, but none are strictly frequency-locked (Bi et al., 2015).
2. Governing Dynamical Models
Bellerophon states have been identified in multiple dynamical models:
Two-group Kuramoto model (Teichmann et al., 2021): with coupling matrix
i.e., attractive intra-group, repulsive inter-group interactions.
General Kuramoto-type with bimodal/frequency-weighted coupling (Bi et al., 2015): where 's are typically drawn from a symmetric (uni-/bi-modal) Lorentzian.
Stuart-Landau amplitude-phase model (Zhang et al., 2019): where is the complex amplitude, and the coupling constant.
Order parameters such as or (amplitude-plus-phase) track the onset and degree of coherence.
3. Analytical and Numerical Onset: Critical Couplings and Bifurcations
The transition to Bellerophon states is critically dependent on system parameters, especially the coupling strength and the structure of the frequency distribution:
- Critical thresholds (Kuramoto-type): For each group with Lorentzian frequency , the critical intra-group coupling for incipient synchronization is . In two-group models, the Bellerophon regime arises for after both groups become partially synchronous (Teichmann et al., 2021).
- Cluster formation: Above this threshold, but before full synchronization (), the system settles into a multi-cluster Bellerophon state where , interpreted as beating mean fields.
- Self-consistency and forward transition (frequency-weighted): The onset in the bimodal Lorentzian scenario is given (upon stability analysis) by
and Bellerophon states appear in the continuous transition regime when , with full synchronization at (Bi et al., 2015).
- Amplitude-phase networks: The structure of the synchronization bifurcation and the presence of a Bellerophon window are controlled by the frequency distribution width (): narrow distributions yield first-order transitions with no Bellerophon state, while broader ones support an intermediate window (Zhang et al., 2019).
4. Cluster Structure, Quantization, and Frequency Profiles
Bellerophon states exhibit quantized frequency clustering and characteristic internal dynamics:
- Staircase quantization: Clusters are characterized by average frequencies
with determined by self-consistent order parameter equations (Bi et al., 2015).
- Cusped profiles: Instantaneous frequencies oscillate around the cluster mean,
with amplitude peaking at cluster edges, producing pronounced cusps.
- Loop quantization: Oscillators in perform revolutions per fundamental period , enforcing equality of average speeds.
- Cluster partitioning in two-group models: The mean field beat, at , creates resonance conditions for cluster formation at
where , (Teichmann et al., 2021).
5. Macroscopic Order Parameters and Dynamical Phase Diagrams
Order parameters provide macroscopic diagnostics of the Bellerophon regime:
- Global order parameter or : In the Bellerophon window, increases continuously but typically oscillates quasiperiodically in time; it sharply rises at the transition to full synchrony.
- Cluster-resolved order parameters : Track coherence within clusters, revealing closed ovals in the complex plane, oscillating at respective cluster’s frequency.
- Bifurcation/phase diagrams: With coupling parameter and frequency distribution as axes, three regimes are observed for large enough frequency width: fully incoherent, intermediate Bellerophon (multi-cluster), and fully synchronized. The transitions change from explosive to continuous as distributional width increases. A summary is presented below for the Stuart-Landau class (Zhang et al., 2019):
| Distribution (width Δ) | IS→BS Critical K | BS sub-transition K | BS→SS Critical K |
|---|---|---|---|
| Lorentzian (Δ=0.5) | ~0.23 | ~1.73 | ~2.13 |
| Triangular (Δ=0.5) | ~0.74 | ~1.22 | ~2.08 |
| Uniform (Δ=0.5) | ~0.64 | ~0.75 | ~2.16 |
Here, IS = incoherent state; BS = Bellerophon state; SS = full synchronization.
6. Internal Dynamics and Two-Stage Structure
In amplitude-phase (Stuart-Landau) networks, the Bellerophon regime itself splits into two sub-stages (Zhang et al., 2019):
- Stage I: Chaotic phase synchronization; phase order but amplitude coherence . Oscillators' trajectories are on chaotic attractors with positive Lyapunov exponent; quantized average-frequency clusters emerge.
- Stage II: Periodic phase synchronization; both and are finite, and amplitudes synchronize onto limit cycles. Time-averaged frequencies remain quantized and instantaneous variations are minimized, with Lyapunov exponent near zero.
This division is determined by the numerical rise in and the Lyapunov exponent's change.
7. Universality, Physical Interpretation, and Experimental Detection
Bellerophon states arise in diverse contexts beyond canonical models:
- Competing mean fields: Whenever two or more mean fields with differing frequency components interact, resonant entrainment can induce Bellerophon structure (Teichmann et al., 2021).
- Physical/metaphorical interpretation: The two-group Kuramoto setting parallels scenarios where subpopulations (e.g., social "parties") have cohesive internal coupling and antagonistic interaction with others; resulting dynamics refuse global consensus but yield multi-cluster opinions.
- Robustness: Bellerophon states have been verified in conformist-contrarian networks, higher-order coupling, and Josephson junction arrays, indicating universal mechanisms tied to nontrivial frequency splitting and mean field competition.
- Experimental detection: Requires global coupling and a sufficiently broad, symmetric frequency distribution; measurement protocols entail mapping instantaneous frequencies versus natural frequencies for quantized steps and cusped internal structure, alongside monitoring and cluster-resolved order parameters (Bi et al., 2015). In amplitude-phase models, detection also involves the observation of transitions within the Bellerophon regime using amplitude and phase order parameters and Lyapunov diagnostics (Zhang et al., 2019).
A plausible implication is that the universality of Bellerophon states may enable their detection across engineered and natural oscillator networks with appropriate coupling and frequency dispersion.
Key references:
- "The Bellerophon state: a novel coherent phase of globally coupled oscillators" (Bi et al., 2015)
- "Partial synchronization in the Kuramoto model with attractive and repulsive interactions via the Bellerophon state" (Teichmann et al., 2021)
- "Novel transition and Bellerophon state in coupled Stuart-Landau oscillators" (Zhang et al., 2019)