Latch-X: Availability & Risk Modeling with Bayesian Networks and Monte Carlo Simulation

Abstract.

Latch-X unifies Fault-Tree intuition, Bayesian-Network (BN) inference, and discrete-event Monte Carlo (MC) simulation into a single declarative YAML language. The framework introduces two practical abstractions: (1) a do: causal relation linking human/automation/disaster risks to technical components, and (2) a latch node, a one-shot, time-bounded transition that captures delayed failover/detection events while preserving BN acyclicity. Latch-X couples these with a schema-validated model, integrity hashblock signing, and a consistent graph specification for reproducible analysis and publication-ready diagrams.

License & scope.

 © 2025 Peter Kováč. Licensed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). See https://creativecommons.org/licenses/by-nc/4.0/ . This license applies to the whitepaper text and figures only; software, source code, and any normative schema/specification are not covered unless expressly stated. “Latch-X” is a registered EU trademark (EUTM No. 019183034); no trademark rights are granted."