Skip to content

Optimizer

pymoo-based multi-objective optimization on trained surrogates.

OptimizerConfig

surrox.OptimizerConfig

Bases: BaseModel

Configuration for the pymoo-based optimizer.

Attributes:

Name Type Description
population_size int

Population size for the evolutionary algorithm.

n_generations int

Number of generations to evolve.

seed int

Random seed for reproducibility.

extrapolation_k int

Number of nearest neighbors for extrapolation detection.

extrapolation_threshold float

Distance threshold beyond which a point is flagged as extrapolating.

constraint_confidence float

Conformal prediction confidence level for constraint evaluation.

See Results for OptimizationResult, EvaluatedPoint, and ConstraintEvaluation.