- examples/animate_propagation_of_chaos.py: 4-panel McKean-Vlasov
simulator at N in {20, 100, 500, 4000} with reference N=12000;
bottom panel tracks W_2(mu^N_t, mu_t) on log scale -> visible
1/sqrt(N) decay (Sznitman 1991).
- examples/propagation_of_chaos.gif (1.6 MB)
- README: new 'Propagation of chaos' subsection under
Mean-field & agent-based dynamics, with empirical-measure
formula, k-tuple factorisation and GIF embed.
- examples/notebooks/14_mckean_vlasov.ipynb: sandwich PRE/code/POST
cells demonstrating W2 ~ 1/sqrt(N) on the same simulator.
Verified executed: sqrt(N)*W2 ~ 0.7 across N (theoretical const).
Notebooks 07 (topology), 08 (volterra), 10 (bsde) and 14 (mckean_vlasov)
now follow the same pedagogical template as the optimal-control tutorial:
- Theorem / proof markdown PRE-cells stating the equation pivot, with
derivations inspired by the latex coursework on path integrals,
Volterra-Malliavin and math-physics-finance lectures.
- Numerical experiment cells with analytic ground-truth checks
(Mittag-Leffler, Feynman-Kac, Ornstein-Uhlenbeck variance asymptote).
- Markdown POST-cells stating the expected result, how to read each
figure, and the conclusion linking back to the API.
- Concrete real-world applications:
* 07 topology -> physics: persistent H1 detects the hole of a thin
annulus vs a filled disk.
* 08 volterra -> sub-diffusion fractional Fokker-Planck moments.
* 10 bsde -> heat equation expectation as a linear BSDE.
* 14 mckean_vlasov -> opinion dynamics on a population.
All cells executed end-to-end with the rhftlab kernel; outputs (figures,
prints, ground-truth errors) are embedded as proof of work.
Includes the deterministic builder script _build_enriched_v2.py used to
regenerate the four notebooks.
Each of the eight v2.0 companion notebooks (10_bsde through
17_generative_calibration) now follows the mandatory pedagogical
sandwich structure:
PRE markdown : theorem / model / pivot equation / what the cell verifies
CODE cell : labelled prints + at least one matplotlib figure
POST markdown: expected result, graph reading, conclusion
Each notebook carries at least one concrete real-world example
(heat plate, inverted pendulum, opinion polarization, collective
decision, OU drift under Cauchy noise, mixture vs gaussian MMD, etc.)
Generator script: scripts/enrich_v2_notebooks.py
Doc plots refreshed via scripts/inject_doc_plots.py.
PyO3 abi3 bindings for the 13 v2.0.0 functions across 8 module groups:
bsde, pde, stochastic_control, optimal_control::quadratic_impact_control,
mean_field::mckean_vlasov, agent_based, inference, optimization.
8 executed companion notebooks under examples/notebooks/10_bsde.ipynb …
17_generative_calibration.ipynb (cell outputs and matplotlib figures
preserved as proof-of-work; verified against analytic ground truths).
8 Sphinx RST pages under docs/source/algorithms/{bsde,pde,stochastic_control,
quadratic_impact_control,mckean_vlasov,agent_based,robust_drift,
generative_calibration_hooks}.rst with .. math:: derivations and inline
.. image:: directives placed immediately after each .. code-block:: python
so each plot appears directly under the code that produced it.
18 PNG plot assets under docs/source/_static/v2/<group>/.
index.rst extended with a new 'v2.0 Generic Stochastic Control & PDE'
toctree caption.
Forbidden-vocabulary audit on new src/, docs/source/algorithms/ and
binding files: zero matches.
All previously stable APIs untouched; v2.0.0 is additive at the binding
level — no v1.x function signature was changed.
ThotDjehuty