ARGOS Hotel Optimization Framework

The repository is organized for easy extension and reproduction of the main experimental results.

ARGOS: Adaptive Recursive Gradient Optimization System for Hierarchical Decision-Making

ARGOS is a research framework that integrates:

• Lexicographic Constraint Optimization (LCO) – strict priority tiers
• Componentwise Approximated Gradient (CAG) – selective Newton / contour filtering

The code implements:

• Lexicographic Constraint Optimization (LCO)
• Componentwise Approximated Gradient (CAG)
• The unified ARGOS operator (LCO + CAG)
• Dynamic hotel management simulation (single-property and multi-property)
• QUBO formulation helpers
• Ablation studies and hyperparameter sensitivity utilities

to produce a stable, dynamic, and lexicographically safe optimization engine for hierarchical decision-making.

The initial application domain is hotel and multi-property management: occupancy control, staffing, fatigue, and revenue optimization under hard service and regulatory constraints. The framework is designed, however, to extend to healthcare, logistics, and quantum / QUBO-based optimization.

This repository contains:

• A Python implementation of ARGOS (LCO + CAG)
• A simulated hotel environment (single-property and multi-unit)
• Ablation experiments (Newton-only vs CAG-only vs full ARGOS)
• Colab-ready notebooks for reproducing the paper’s results

---

1. Repository Structure

argos_hotel_optimization/
├── src/
│   └── argos/
│       ├── __init__.py
│       ├── lco.py           # Lexicographic tier definitions & projections
│       ├── cag.py           # Componentwise Approximated Gradient filter
│       ├── argos_core.py    # Unified ARGOS optimizer (LCO + CAG)
│       ├── hotel_env.py     # Single-unit hotel CMDP-style environment
│       ├── multiunit.py     # Multi-property wrapper around hotel_env + ARGOS
│       └── qubo.py          # QUBO helper stubs / placeholders
│
├── experiments/
│   ├── run_single_hotel.py  # Main single-hotel experiment script
│   ├── run_multi_unit.py    # Multi-unit experiment script
│   └── run_ablation.py      # Newton-only vs CAG-only vs ARGOS comparison
│
├── notebooks/
│   ├── ARGOS_Single_Hotel_Demo.ipynb
│   ├── ARGOS_Multi_Unit_Demo.ipynb
│   └── ARGOS_Ablation_Study.ipynb
│
├── data/
│   ├── synthetic_single_hotel.csv  # Example state / metric traces
│   └── synthetic_multi_hotel.csv   # Example multi-unit traces
│
├── docs/
│   ├── INSTALL.md
│   └── USAGE.md
│
├── README.md
├── requirements.txt
└── LICENSE

Data Availability

All datasets used in the ARGOS experiments are fully synthetic and openly available on Zenodo:

Dataset DOI:
https://doi.org/10.5281/zenodo.17645085

Software Metadata

Name: ARGOS — Adaptive Recursive Gradient Optimization System
Version: 1.0.1
Release Date: 2025
License: MIT
Repository: https://github.com/Galactic-Code-Developers/Argos-Hotel-Optimization
Programming Language: Python 3.9+
Primary Dependencies: NumPy, Pandas, Matplotlib
Supported Platforms: Linux, macOS, Windows
Continuous Integration: GitHub Actions (optional)
Documentation: Included in docs/ and notebooks in notebooks/

Primary Purpose:
Stable hierarchical optimization under strict lexicographic priorities, combining
Lexicographic Constraint Optimization (LCO) with Componentwise Approximated Gradient (CAG).

Research Domains:

• Operations Research
• Optimization & Control
• Reinforcement Learning (CMDP-style)
• Hospitality Management Systems
• Multi-agent & Multi-unit resource allocation

Key Features:

• Lexicographically safe updates (Tier-1 invariants always preserved)
• Componentwise selective gradient filtering (CAG)
• Integrated LCO + CAG update engine (ARGOS Core)
• Single-unit and multi-unit hotel environment simulators
• Ablation tools (Newton-only, CAG-only, full ARGOS)
• Reproducible experiments via CLI and Colab notebooks

Intended Users:
Researchers, operations analysts, optimization practitioners, and academic collaborators evaluating lexicographically constrained decision systems.

How to Cite

If you use the ARGOS optimization system or its accompanying synthetic datasets, please cite both the dataset and the software using the references below.

---

🔹 Dataset Citation

Valamontes, A. (2025). ARGOS Synthetic Hotel Optimization Datasets [Data set].
Kapodistrian Academy of Science.
https://doi.org/10.5281/zenodo.17645086

---

🔹 Software Citation (ARGOS System)

Valamontes, A., & Research Team, C. (2025). ARGOS: Adaptive Recursive Gradient Optimization System (Version 1.0.1).
Kapodistrian Academy of Science.
https://doi.org/10.5281/zenodo.17645169

---

🔹 BibTeX Entries

@dataset{Valamontes_ARGOS_Datasets_2025,
  author       = {Valamontes, Antonios},
  title        = {ARGOS Synthetic Hotel Optimization Datasets},
  year         = {2025},
  publisher    = {Kapodistrian Academy of Science},
  doi          = {10.5281/zenodo.17645086},
  url          = {https://doi.org/10.5281/zenodo.17645086},
  version      = {1.0.0},
  note         = {Data set}
}

@software{Valamontes_ARGOS_2025,
  author       = {Valamontes, Antonios and Research Team, C.},
  title        = {ARGOS: Adaptive Recursive Gradient Optimization System},
  year         = {2025},
  publisher    = {Kapodistrian Academy of Science},
  doi          = {10.5281/zenodo.17645169},
  url          = {https://doi.org/10.5281/zenodo.17645169},
  version      = {1.0.1}
}