Skip to content

ARGOS: Adaptive Recursive Gradient Optimization System v1.0.1

Latest

Choose a tag to compare

@geopayme geopayme released this 19 Nov 02:21
· 8 commits to main since this release
Immutable release. Only release title and notes can be modified.
e8e0966

Software Metadata

Name: ARGOS — Adaptive Recursive Gradient Optimization System
Version: 1.0.1
Release Date: 2025
License: MIT
Repository: https://github.com//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:
Valamontes, A. (2025). ARGOS: Adaptive Recursive Gradient Optimization System (Preprint).

Full Changelog: ARGOSv1...ARGOSv2