Problem#

This class can be used to define a rare event estimation problem, which can be passed to the solvers of this package.

class rareeventestimation.NormalProblem(lsf: Callable, dim: int, sample_size: int, prob_fail_true=None, mpp=None, hints=None, name=None)#

This class defines a special kind of rare event estimation problem. It assumes that the lsf operates on the standard normal space. Thus the energy function and the sample distrubtion are known (standard normal).

Constructor if distribution is standard normal.

Parameters:

  • lsf (Callable) – Handle to limit state function.

  • dim (int) – Dimension of the multivariate normal distribution.

  • sample_size (int) – Initial sample size.

  • prob_fail_true (optional) – True probability of failure. Defaults to None. Can be used for performance evaluations in postprocessing.

  • mpp (optional) – Most probable point of failure. Defaults to None. Can be used for performance evaluations in postprocessing.

  • hints (dict, optional) – Dictionary with keyword arguments for solver.

  • name (str,optional) – Name of problem for stringification.