Published June 5, 2026
| Version v1.0
Software
Open
PyBELSA v1.0: Bayesian Epistemic Logic for Sustainable Aquifers — reproducibility archive
Description
Python implementation and reproducibility archive for the paper: 'An empirical crossover law for model-agnostic POMDP policies under transition-model misspecification' (Jakaite & Schetinin, 2026, Nature Machine Intelligence, under review).
Includes: BELSA framework (ELS_Phil_T, ELS_Int, ELS_Adapt), all ten benchmark environments (B1–B10), pre-computed Monte Carlo results (N_MC=200, SEED=2024), publication figures (Fig. 1–4), and a one-command reproduction script (reproduce_b10.py) for the headline result.
Notes
Verified on Python 3.13, numpy 2.4.3, scipy 1.17.1, matplotlib 3.10.8. All pre-computed outputs included; no re-running required to reproduce paper figures.
Files
vitsch/PyBELSA-v1.0.zip
Files
(3.7 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:38af0a19321506c2aade7d442313d5c7
|
3.7 MB | Preview Download |
Additional details
Related works
- Is supplement to
- Preprint: 10.2139/ssrn.6308238 (DOI)
- Is version of
- https://github.com/vitsch/PyBELSA (URL)
Software
- Repository URL
- https://github.com/vitsch/PyBELSA