opendose_poppk.population

Population simulation with covariates.

Classes

PopulationSimulator : Monte Carlo simulation of heterogeneous populations

class opendose_poppk.population.PopulationSimulator(pk: PKModel | None = None, pd: PDModel | None = None, covariate_model: CovariateModel | None = None, dose: float = 1000.0)

Bases: object

Monte Carlo simulation of heterogeneous population with covariates.

Integrates PKModel + PDModel + CovariateModel to generate individual PK/PD profiles with realistic covariate distributions.

Example

>>> pk  = PKModel(F=0.8, ka=1.8, ke=0.28, Vd=65)
>>> pd  = PDModel(EC50=10.0, n=1.5)
>>> sim = PopulationSimulator(pk, pd, dose=1000.0)
>>> res = sim.run(n_subjects=1000, t_max=12.0,
...               covariates={"weight": ("normal", 70, 15),
...                           "crcl":   ("normal", 90, 30)})

run(n_subjects: int = 1000, t_max: float = 24.0, n_points: int = 200, covariates: dict | None = None, seed: int = 42) → dict

Execute Monte Carlo simulation.

Parameters:

n_subjects (simulated population size)
t_max (final time (h))
n_points (number of points in time curve)
covariates (covariate distributions.) – Format: {“name”: (“type”, p1, p2)} Types: “normal”, “uniform”, “lognormal” Example: {“weight”: (“normal”, 70, 15)}
seed (random seed)

Returns:

percentiles_pd, covariates_sim

Return type:

dict with t, pk_profiles, pd_profiles, percentiles_pk,

simulate(n: int = 10) → list[ndarray]

Legacy compatibility method: returns list of profiles.

This simplifies testing - just check the array size.