Skip to documentation content

Constraint penalty

Adjust constraint penalty when configuring a CMA-ES search.

Use constraintPenalty to set this option for CMA-ES. The initial value is 1000.0. Start with a valid solved model and change optimization controls deliberately, comparing feasibility and convergence with the previous run. A tighter tolerance or a larger budget does not compensate for an incorrect objective or invalid design bounds.

Constraint penalty — Telion Python API 0.1.0-alpha | Telion