calibr8.optimization

The optimization module implements convenience functions for maximum likelihood estimation of calibration model parameters.

calibr8.optimization.fit_scipy(model: CalibrationModel, *, independent: ndarray, dependent: ndarray, theta_guess: Sequence[float] | ndarray, theta_bounds: Sequence[Tuple[float, float]], minimize_kwargs: Mapping[str, Any] | None = None)

Function to fit the calibration model with observed data.

Parameters:

modelcalibr8.CalibrationModel: the calibration model to fit (inplace)
independentarray-like: desired values of the independent variable or measured values of the same
dependentarray-like: observations of dependent variable
theta_guessarray-like: initial guess for parameters describing the PDF of the dependent variable
theta_boundsarray-like: bounds to fit the parameters
minimize_kwargsdict: keyword-arguments for scipy.optimize.minimize

Returns:

thetaarray-like: best found parameter vector
historylist: history of the optimization

calibr8.optimization.fit_scipy_global(model: CalibrationModel, *, independent: ndarray, dependent: ndarray, theta_bounds: list, method: str | None = None, maxiter: int = 5000, minimizer_kwargs: Mapping[str, Any] | None = None)

Function to fit the calibration model with observed data using global optimization.

Parameters:

modelcalibr8.CalibrationModel

the calibration model to fit (inplace)

independentarray-like

desired values of the independent variable or measured values of the same

dependentarray-like

observations of dependent variable

theta_boundsarray-like

bounds to fit the parameters

method: str, optional

Type of solver. Must be one of the following:

"dual_annealing"

If not given, defaults to "dual_annealing".

maxiter: int, optional

Maximum number of iterations of the dual_annealing solver. If not given, defaults to 5000.

minimize_kwargsdict

keyword-arguments for scipy.optimize.minimize

Returns:

thetaarray-like: best found parameter vector
historylist: history of the optimization, containing best parameter, objective value, and solver status

Raises:

ValueError: user input does not match number of parameters
ValueError: user specifies not supported optimization method
Warning: fit failed or bounds constrain optimization