Namespace OptunaSharp.Samplers.AgeMoea

Classes

AgeMoea2ElitePopulationSelectionStrategy

Elite-population selection for AgeMoea2Sampler: the AGE-MOEA-2 environmental selection (OptunaSharp.Samplers.AgeMoea.AgeMoea2Survival) applied to Optuna trials. Objective values are sign-adjusted to minimization (like _rank_population), fed through the survival geometry, and the survivors are returned ordered by front rank (ascending) then AGE-MOEA-2 crowding / survival score (descending) — so the extreme and most-spread solutions seed the parent population first.

This is the AGE-MOEA-2-specific piece of Layer 2. Child generation stays Optuna-idiomatic (SBX/PM over the transformed search space), so the sampler handles Optuna's full parameter space; the byte-perfect pymoo-faithful AGE-MOEA-2 is Layer 1 (OptunaSharp.Samplers.AgeMoea.AgeMoea2Optimizer). Constrained optimization is not yet modeled here (unconstrained multi-objective only).