🔴 ENTRO-ENGINE — Entropy Flow Regulator for Coupled Intelligence Systems
"A single engine regulates itself. A fleet of engines requires a coordinator." — Samir Baladi, April 2026
ENTRO-ENGINE is the fourth project of the EntropyLab research program (E-LAB-04).
It advances from single-system entropy control to multi-system entropy coordination — distributing thermodynamic pressure across coupled subsystems under a shared global budget constraint.
📌 DOI & Badges
🌐 Official Website
https://entro-engine.netlify.app
🎯 Vision
Build a coordination layer for AI pipelines in which multiple entropy-generating subsystems share a common thermodynamic budget — ensuring that no single subsystem collapses under load while others operate at idle, and that corrective pressure is distributed in proportion to each subsystem's entropy contribution.
The Core Shift
| Architecture | Scope | Objective |
|---|---|---|
| ENTROPIA (E-LAB-01) | Definition | Establish Ψ vocabulary |
| ENTRO-AI (E-LAB-02) | Measurement | Monitor & detect collapse |
| ENTRO-CORE (E-LAB-03) | Local Control | Stabilize single Ψ field |
| ENTRO-ENGINE (E-LAB-04) | Coordination | Distribute pressure across coupled systems |
🔬 Key Innovation
Instead of: Each subsystem independently minimizes its own entropy
ENTRO-ENGINE does: Enforce a global entropy budget across all coupled subsystems
Ψ_total(t) = Ψ_scraper + Ψ_llm ≤ Ψ_budget ∀t
The engine computes corrective signals proportional to each subsystem's entropy fraction:
u_i(t) = −α · ρ_i(t) · tanh(β · Δ(t))
Where:
ρ_i= fractional entropy ratio (fairness principle)Δ(t)= global entropy pressure above budgettanh= saturation (prevents overcorrection)
📊 Architecture
entro-engine/
│
├── engine/ # Core coordination engine
│ ├── controller.py # ENTRO-ENGINE control law (Eq. 14)
│ ├── budget.py # Entropy budget constraint (Eq. 9–11)
│ └── allocator.py # Proportional allocation ratios (Eq. 12–13)
│
├── subsystems/ # Coupled subsystem models
│ ├── scraper.py # Scraper entropy source (Eq. 1, 4, 6)
│ ├── llm_queue.py # LLM processing queue (Eq. 2, 5, 7)
│ └── base.py # Abstract subsystem interface
│
├── entropy/ # Entropy formulation layer
│ ├── proxy.py # Log-proxy functions: Ψ = log(1 + x)
│ ├── aggregator.py # Global Ψ_total computation (Eq. 8)
│ └── normalizer.py # Logistic normalization
│
├── pipeline/ # Realistic pipeline simulation
│ ├── simulator.py # Three-phase pipeline (warmup / surge / recovery)
│ ├── simulator_v9.py # Balanced regime engine
│ └── simulator_v10.py # Recursive attractor engine
│
├── reports/ # Daily/Weekly/Monthly reports
│ ├── daily/ # Daily .txt reports
│ ├── weekly/ # Weekly summaries
│ ├── monthly/ # Monthly analyses
│ └── alerts/ # Alert logs
│
├── tests/ # Unit & integration tests
│ ├── test_controller.py # Controller tests (9/9 passed)
│ ├── test_allocator.py # Allocator tests
│ └── test_integration.py # Integration tests
│
├── data/ # Simulation output datasets
│ └── simulation_runs/ # CSV trajectory files
│
├── docs/ # Documentation
├── paper/ # Research paper assets
├── AUTHORS.md
├── CHANGELOG.md
├── LICENSE
├── README.md
└── pyproject.toml
⚙️ Core Equations
| Eq. | Expression | Description |
|---|---|---|
| 6 | Ψ_s(t) = log(1 + r_s(t)) |
Scraper entropy proxy |
| 7 | Ψ_l(t) = log(1 + q_l(t)) |
LLM entropy proxy |
| 8 | Ψ_total(t) = Ψ_s + Ψ_l |
Global entropy aggregate |
| 9 | Ψ_total(t) ≤ Ψ_budget ∀t |
Budget constraint |
| 11 | Δ(t) = Ψ_total(t) − Ψ_budget |
Entropy pressure |
| 12–13 | ρ_i = Ψ_i / Ψ_total |
Proportional allocation ratios |
| 14 | u_i(t) = −α · ρ_i · tanh(β · Δ) |
ENTRO-ENGINE control law |
📦 Installation
pip install entro-engine
From source:
git clone https://gitlab.com/gitdeeper10/entro-engine.git
cd entro-engine
pip install -e .
🚀 Quick Start
from entro_engine import ENTROEngineController
# Initialize controller
controller = ENTROEngineController(alpha=0.5, beta=2.0, psi_budget=3.0)
# Compute control signal
signal = controller.compute_control(psi_i=1.5, psi_total=4.0)
print(f"u = {signal.u:.3f}")
print(f"ρ = {signal.rho:.3f}")
print(f"Δ = {signal.delta:.3f}")
📈 Simulation Results
Validated on a three-phase realistic pipeline (warmup → surge → recovery):
Metric Uncontrolled ENTRO-ENGINE Improvement Budget violations 312 / 1000 12 / 1000 96.2% reduction Peak Ψ_total 5.84 3.31 43.3% reduction Peak LLM queue 87.6 req 23.4 req 73.3% reduction
Regime Distribution (v9 Final Run)
Regime Count Percentage 🟢 Stable 24 30.0% 🟡 Stress 19 23.75% 🔴 Critical 37 46.25%
🔗 Links
Resource URL 🌐 Official Website https://entro-engine.netlify.app 📊 Dashboard https://entro-engine.netlify.app/dashboard 📚 Documentation https://entro-engine.netlify.app/documentation 📋 Reports https://entro-engine.netlify.app/reports 🦊 GitLab (Primary) https://gitlab.com/gitdeeper10/entro-engine 💻 GitHub (Mirror) https://github.com/gitdeeper10/entro-engine 📦 PyPI https://pypi.org/project/entro-engine 📌 DOI https://doi.org/10.5281/zenodo.19441032 🔴 ENTROPIA (E-LAB-01) https://doi.org/10.5281/zenodo.19416737 ⚙️ ENTRO-AI (E-LAB-02) https://doi.org/10.5281/zenodo.19284086 🧠 ENTRO-CORE (E-LAB-03) https://doi.org/10.5281/zenodo.19431029 👤 ORCID https://orcid.org/0009-0003-8903-0029
🗺️ EntropyLab Research Program
Project Title Status DOI E-LAB-01 ENTROPIA: Thermodynamic-Information Unification ✅ Published 10.5281/zenodo.19416737 E-LAB-02 ENTRO-AI: Entropy-Resistant AI Architectures ✅ Published 10.5281/zenodo.19284086 E-LAB-03 ENTRO-CORE: Self-Regulated Intelligence ✅ Published 10.5281/zenodo.19431029 E-LAB-04 ENTRO-ENGINE: Entropy Flow Regulator ✅ Published 10.5281/zenodo.19441032 E-LAB-05 ENTRO-FIN: Financial Market Entropic Signatures 📋 Planned — E-LAB-06 ENTRO-SOCIAL: Social Network Information Cascades 📋 Planned — E-LAB-07 ENTRO-QUANTUM: Quantum Extension 📋 Planned — E-LAB-08 ENTRO-BIO: Biological Neural Networks 📋 Planned — E-LAB-09 ENTRO-MANIFESTO: Unified Research Manifesto 📋 Planned —
📄 Citation
@software{baladi2026entroengine,
author = {Baladi, Samir},
title = {ENTRO-ENGINE: Entropy Flow Regulator for Coupled Intelligence Systems},
year = {2026},
version = {1.0.0},
doi = {10.5281/zenodo.19441032},
url = {https://gitlab.com/gitdeeper10/entro-engine},
note = {E-LAB-04. Builds on E-LAB-01, E-LAB-02, E-LAB-03}
}
📜 License
MIT License — Copyright © 2026 Samir Baladi
Part of the EntropyLab ten-project research program "Intelligence by Design, Stability by Physics"
Metadata
Release files for entro-engine 1.0.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| entro_engine-1.0.0.tar.gz | 20.9 kB | Without provenance | Details for entro_engine-1.0.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| entro_engine-1.0.0-py3-none-any.whl | Python 3 | none | any | Details for entro_engine-1.0.0-py3-none-any.whl |
Total release size: 29.4 kB
Release files / entro_engine-1.0.0.tar.gz
| Download URL | entro_engine-1.0.0.tar.gz |
|---|---|
| Size | 20.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
34443d7f25158090332a666be918f5004be7e204b7b66d345785394387666f6f
|
|
BLAKE2b-256 checksum How to use checksums |
0a66ecd74777ac1038d27a9498ed21853a9db349de0ee818ab6f5809b6a50531
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
ENTRO-ENGINE-Uploader/1.0
|
Release files / entro_engine-1.0.0-py3-none-any.whl
| Download URL | entro_engine-1.0.0-py3-none-any.whl |
|---|---|
| Size | 8.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
319260710b74b9d985dddeaaf55c09da680c38b10a02fc1c3834fb14bfb20963
|
|
BLAKE2b-256 checksum How to use checksums |
1ffcaafecba08d8f042c8258867c02fc9df0fb81a9bfcbbb408a68151b798164
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
ENTRO-ENGINE-Uploader/1.0
|
Security for entro-engine 1.0.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.