ThotDjehuty
d8682f61e5
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.
73 lines
2.6 KiB
ReStructuredText
73 lines
2.6 KiB
ReStructuredText
McKean–Vlasov — propagation of chaos
|
||
====================================
|
||
|
||
Interacting-particle Euler scheme for $dX_t = θ(\bar X_t - X_t) dt + σ dW_t$ (`mean_reverting_mckean_vlasov`). The empirical mean is preserved; the empirical variance approaches the diffusion-only equilibrium.
|
||
|
||
.. note:: Companion executed notebook: `14_mckean_vlasov.ipynb <../../examples/notebooks/14_mckean_vlasov.ipynb>`_
|
||
|
||
14 — McKean–Vlasov mean-reverting dynamics
|
||
==========================================
|
||
|
||
.. code-block:: python
|
||
|
||
import numpy as np
|
||
import matplotlib.pyplot as plt
|
||
from optimizr import _core as opt
|
||
plt.rcParams['figure.figsize'] = (7, 4)
|
||
plt.rcParams['figure.dpi'] = 110
|
||
|
||
.. code-block:: python
|
||
|
||
init = np.linspace(-2.0, 2.0, 200).tolist()
|
||
init_mean = float(np.mean(init))
|
||
res = opt.mean_reverting_mckean_vlasov(
|
||
initial=init, theta=1.0, sigma=0.1,
|
||
n_steps=1000, t_horizon=1.0, seed=42,
|
||
)
|
||
n_t = res['n_steps']; n_p = res['n_particles']
|
||
X = np.array(res['paths_flat']).reshape(n_t, n_p)
|
||
tg = np.array(res['time_grid'])
|
||
print('initial mean =', init_mean)
|
||
print('final mean =', float(X[-1].mean()))
|
||
print('final std =', float(X[-1].std()))
|
||
|
||
.. code-block:: python
|
||
|
||
fig, ax = plt.subplots()
|
||
ax.plot(tg, X[:, ::20], color='tab:blue', alpha=0.2, lw=0.6)
|
||
ax.plot(tg, X.mean(axis=1), color='red', lw=2, label='empirical mean')
|
||
ax.axhline(init_mean, color='k', ls=':', label='initial mean')
|
||
ax.set_xlabel('t'); ax.set_ylabel('X^i_t'); ax.legend(); ax.grid(alpha=0.3)
|
||
ax.set_title('Mean-reverting McKean–Vlasov — 200 particles')
|
||
fig.tight_layout(); plt.show()
|
||
|
||
.. image:: ../_static/v2/mckean_vlasov/plot_01.png
|
||
:align: center
|
||
:width: 80%
|
||
|
||
.. code-block:: python
|
||
|
||
fig, ax = plt.subplots()
|
||
ax.hist(X[0], bins=30, alpha=0.5, label='t = 0', density=True)
|
||
ax.hist(X[-1], bins=30, alpha=0.5, label='t = T', density=True)
|
||
ax.set_xlabel('x'); ax.set_ylabel('empirical density'); ax.legend(); ax.grid(alpha=0.3)
|
||
ax.set_title('Marginal density at t = 0 and t = T')
|
||
fig.tight_layout(); plt.show()
|
||
|
||
.. image:: ../_static/v2/mckean_vlasov/plot_02.png
|
||
:align: center
|
||
:width: 80%
|
||
|
||
**Verified:** empirical mean stays within `0.05` of the initial mean.
|
||
|
||
API
|
||
---
|
||
|
||
.. code-block:: rust
|
||
|
||
pub fn simulate_mckean_vlasov<B>(initial: &[f64], drift: B, cfg: &McKeanVlasovConfig) -> Result<McKeanVlasovResult>
|
||
where B: Fn(f64, &[f64]) -> f64;
|
||
|
||
pub struct McKeanVlasovConfig { pub n_particles: usize, pub n_steps: usize, pub t_horizon: f64, pub sigma: f64, pub seed: u64 }
|
||
pub struct McKeanVlasovResult { pub paths: Array2<f64>, pub time_grid: Array1<f64> }
|