GEPA.Strategies.CandidateSelector.Pareto (GEPA v0.3.0)

Copy Markdown View Source

Selects candidates from Pareto front using frequency-weighted sampling.

Programs appearing in more Pareto fronts have higher selection probability. This naturally balances specialization (programs good at specific examples) with generalization (programs good across many examples).

Summary

Functions

select(state, rand_state)

@spec select(GEPA.State.t(), :rand.state() | nil) ::
  {GEPA.Types.program_idx(), :rand.state()}