Skip to content

ENAQT Optimal-Noise Scan

Abstract

BL-87 provides a bounded local simulator for environment-assisted quantum transport (ENAQT). It scans a finite set of local dephasing rates and maximises the population irreversibly transferred into a target sink by a fixed time. The committed evidence contains one disordered-chain intermediate optimum and two negative controls. This is scenario-specific transport evidence, not a universal noise optimum or a synchronisation, biological, hardware, or consciousness result.

Introduction and literature boundary

Plenio and Huelga showed that local dephasing can improve excitation transport in selected dissipative quantum networks. Mohseni and co-authors studied the corresponding environment-assisted transfer mechanism in a model of the FMO photosynthetic complex. Those results motivate a transport-efficiency scan; they do not license replacing efficiency with a phase estimator or transferring the conclusion to Kuramoto synchronisation, BKT physics, consciousness, or a physical noise-control policy.

Methodology

The model uses a single-excitation site basis augmented by orthogonal sink and loss states. For a real symmetric hopping matrix K and site energies ω, the network Hamiltonian is

\[ H = \operatorname{diag}(\omega) + K. \]

Each site has a local dephasing jump operator with rate γ. The target has an irreversible sink jump with rate κ, and every site may recombine into a loss state with rate μ. The full trace-preserving Lindblad equation is propagated with the exponential action of a matrix-free generator.

For a chosen horizon T, transfer efficiency is the final sink population

\[ \eta(\gamma;T)=\langle s\rvert\rho_\gamma(T)\lvert s\rangle. \]

The scanner reports the best sampled γ, the exact zero-dephasing endpoint, the largest-scanned-γ endpoint, and whether the best grid point is strictly interior and exceeds both endpoints by a configured minimum. The largest finite dephasing rate is deliberately called “high noise,” not a classical limit.

import numpy as np

from scpn_quantum_control.analysis import enaqt_scan

K = np.zeros((4, 4), dtype=np.float64)
for site in range(3):
    K[site, site + 1] = K[site + 1, site] = 1.0

result = enaqt_scan(
    K,
    np.array([0.0, 3.0, -2.0, 1.0]),
    gamma_range=np.array([0.0, 0.01, 0.03, 0.1, 0.3, 1.0, 3.0, 10.0, 30.0]),
    t_evolve=10.0,
)

assert result.has_intermediate_optimum
assert result.optimal_gamma == 3.0

optimal_r, r_values, coherent_r, and classical_r remain read-only compatibility aliases. They now return transport-efficiency values; the classical_r name does not assert that the finite high-noise endpoint is a classical limit. New code should use the explicit efficiency fields.

Results

Regenerate the committed JSON and Markdown evidence with:

PYTHONPATH=src:oscillatools/src python scripts/run_enaqt_evidence.py
Scenario γ* η(0) η(γ*) High-noise η Ratio Interior?
Disordered four-site chain 3 0.0522739 0.1765650 0.0114666 3.37769 yes
Uniform three-site chain 0 0.8195421 0.8195421 0.0714405 1 no
Disconnected target 0 0 0 0 0 no

The first row demonstrates the intended intermediate-noise effect on one frozen finite model. The second shows that dephasing can be strictly detrimental. The third checks that dephasing does not create a transport path absent from the Hamiltonian. Every row is replayed and digest-bound in data/enaqt_product/bl87_enaqt_evidence.json.

Conclusion and control boundary

BL-87 closes a local simulator and evidence lane. It does not expose a noise setpoint controller: an optimum depends on the network, horizon, sink/loss rates, dephasing model, and sampled grid. Using a simulated γ* to alter a provider, QPU, laboratory noise source, biological system, or plant requires a separately authorised and calibrated control protocol. No such protocol is implemented here.

BL-50's QFI/QNG geometry and BL-60's future chimera/multiscale synchronisation targets remain separate scientific lanes. This transport scan neither consumes their observables nor promotes a transport optimum into a geometry or synchronisation result.