E-LAB-09 · EntropyLab · April 2026

ENTRO-PULSE

Periodic Entropy Pulsing and Informational Wave Management in High-Velocity AI Systems.
From continuous suppression to rhythmic oscillatory control.

δ(t) · EPWM Duty Cycle · Adaptive pulse width modulation

0.62

⟨δ⟩ (duty cycle expectation)

PULSE · Active phase · 38.7% throughput gain

GitHub Repository DOI: 10.5281/zenodo.19547863

01 · Core Innovation

EPWM · RRL · PGC

ENTRO-PULSE introduces Periodic Entropy Pulsing (PEP) — a control paradigm that transforms entropy flow management from continuous suppression into a rhythmically-managed oscillatory regime.

Entropic Pulse Function · Eq 3.1

S_pulse(t; ω, δ) = H(δ - frac(ω·t/2π))

Periodic binary gating signal

EPWM Adaptive Duty Cycle · Eq 4.1

δ(t) = δ_max · exp(-Ψ(t) / (θ_crit - Ψ(t)))

Adaptive duty cycle contracts as Ψ → θ_crit

Rhythmic Resonance Law · Eq 5.1

dφ_i/dt = ω_i - (K/N)·Σ_j sin(φ_i - φ_j - Δφ_target_ij)

Modified Kuramoto for anti-phase synchronization

Pulse-Ghost Control Law · Eq 6.1

u_PGC(t) = S_pulse·[u_base + ζ·(Ψ* - Γ)] + (1-S_pulse)·u_cooldown

Integration with ENTRO-GHOST (E-LAB-08)

02 · Validation Results

Throughput Gain · 38.7% Target Exceeded

Configuration Throughput Collapse Rate Status

Baseline (continuous) 1.000 23.4% Memoryless

EPWM only (ω=0.8) 1.312 4.1% Pulsing active

EPWM + RRL (8 agents) 1.361 1.2% Anti-phase sync

Full PGC 1.387 0.0% ✅ TARGET EXCEEDED

Entropy Pulse Width Modulation (EPWM)

δ(t) = δ_max·exp(-Ψ/(θ_crit-Ψ))

ω (frequency)

0.8 rad/cycle

δ_max (max duty)

0.7

θ_crit (threshold)

0.85

β_s (sharpness)

10.0

Rhythmic Resonance Law (RRL)

r = |(1/N)·Σ exp(i·(φ_j-φ_target_j))|

Order Parameter r

0.08

Peak Load Reduction

86.1%

Coupling Strength K

0.5

Agents (N)

Pulse-Ghost Controller (PGC)

ζ_phase = ζ·(1 + ρ·(1-frac/δ))

α (ghost decay)

0.1

ζ (ghost pull)

0.65

ρ (phase awareness)

0.4

Throughput Gain

38.7%

# pip install entro-pulse
from entro_pulse import EntropyPulseController

epwm = EntropyPulseController(omega=0.8, delta_max=0.7)

# Execute control step
result = epwm.step(psi=0.85, u_base=0.1)

# → Output
Duty cycle: 0.623 · Output: 0.062 · Active: True
Throughput gain: 38.7% ✅ · Zero collapse events

"A system that pulses does not merely survive its entropy dynamics; it dances with them.
The heart does not beat continuously — and neither should an intelligence that wishes to endure."

— Samir Baladi · ENTRO-PULSE · April 2026