Skip to content

Regime Manager

Regimes

Regime Description
NOMINAL Healthy operation. R_good above threshold, no violations.
DEGRADED R_good below target but above critical. Soft warnings possible.
CRITICAL Hard boundary violations or R below critical threshold.
RECOVERY Transitioning from CRITICAL back toward NOMINAL.

Entry Conditions

Regime Condition
NOMINAL avg_R >= 0.6 and no hard violations
DEGRADED 0.3 <= avg_R < 0.6 and no hard violations
CRITICAL avg_R < 0.3 OR hard boundary violations
RECOVERY Previously CRITICAL, now avg_R >= 0.3

avg_R is the mean R across all layers in UPDEState.

Thresholds

See ASSUMPTIONS.md § Regime Thresholds for provenance.

_R_CRITICAL = 0.3
_R_DEGRADED = 0.6

Hysteresis

RegimeManager(hysteresis=0.05, cooldown_steps=10).

  • hysteresis: Not currently used for threshold offset; reserved for future refinement.
  • cooldown_steps: Minimum steps between regime transitions. Prevents oscillation between adjacent regimes.

Exception: escalation to CRITICAL always bypasses cooldown.

Transition Logic

proposed = evaluate(upde_state, boundary_state)
regime   = transition(current, proposed)

transition() applies cooldown gating:

  1. If proposed == current, no change.
  2. If steps_since_last_transition < cooldown_steps AND proposed != CRITICAL, reject transition.
  3. Otherwise, accept transition and reset counter.

Supervisor Integration

SupervisorPolicy.decide() calls RegimeManager.evaluate() and transition() internally, then selects control actions based on the resulting regime:

Regime Action
NOMINAL No-op
DEGRADED Boost global K
CRITICAL Increase zeta, reduce K on worst layer
RECOVERY Gradual K restore

References

  • [acebron2005] J. A. Acebrón et al. (2005). The Kuramoto model: a simple paradigm for synchronization phenomena. Rev. Mod. Phys. 77, 137–185. — Incoherence / partial-sync boundaries motivating R thresholds.
  • [dorfler2014] F. Dörfler & F. Bullo (2014). Synchronization in complex networks of phase oscillators: a survey. Automatica 50, 1539–1564. — Finite-N synchronisation conditions and hysteresis.

Why this manager exists

  • The manager converts noisy numeric signals into bounded, auditable control states.
  • Cooldown is a practical anti-oscillation control that avoids rapid action churn on measurement noise.
  • CRITICAL bypass keeps safety priority explicit: hard faults move the system into protective behavior without waiting for cool-off cycles.

Incident response interpretation

Regime state is what allows teams to treat synchronization as an operations signal. Because transitions are gated, this manager protects against noisy oscillations creating unnecessary action churn.

Typical runbook mapping: - NOMINAL: monitor for drift only; policy remains passive. - DEGRADED: allow preparatory control actions while preserving normal posture. - CRITICAL: enforce hard safety behavior immediately. - RECOVERY: phase actions back into stable trajectory with bounded restoration.

This mapping is intentionally compact and deterministic so operators can reason about expected state behavior from one run to the next.

Control implications for operators

Treat regime transitions as a first-order operating model for action timing:

  • keep cooldown_steps enabled for non-critical transitions to avoid state flip-flop,
  • keep the CRITICAL bypass in place so safety transitions are not delayed,
  • include regime trace in post-run evidence when tuning controllers.

This mapping is the main guard against control chatter in noisy telemetry periods.