Repository navigation
Releases
Releases · nagarjunak-pixel/chronos-mesh
ChronosMesh v1.0.0 — Neuro-Symbolic Supply-Chain Digital Twin & Exact MILP Optimizer
ChronosMesh v1.0.0
Highlights
-
Pure-Python Exact Two-Phase Simplex + Best-First Branch-and-Bound MILP Engine: Bland's anti-cycling rule + min-heap B&B with zero-gap global optimality certificates (~3-12 ms solve latency), verified via Hypothesis against an exhaustive
$2^m$ enumeration oracle. - 4-Tier Diamond BOM Graph & MIT Simchi-Levi Contagion: Recursive multi-echelon component explosion, joint demand/lead-time safety stock, and Time-to-Recovery (TTR) vs. Time-to-Survive (TTS) Revenue-at-Risk quantification.
- 3-Objective Pareto Frontier: Fixed-charge Big-M multi-sourcing with 65% anti-concentration diversification cap across Cost, Lead-Time SLA, and Scope-3 Carbon regimes.
- Monte Carlo Tail-Risk Engine: 500-2,000 trial stochastic shock simulation proving 81.4% CVaR95 reduction (-$937K tail loss) and 100% fill-rate vs. single-sourced baselines.
- Control Tower UI & Verification: FastAPI REST service, Rich CLI, TypeScript/React operational console, strict mypy, and 98.37% pytest coverage.