Response Optimization
Combines several saved regression models and searches for the factor point with the highest joint desirability (Derringer-Suich, multi-start Nelder-Mead).
Response Optimization
The response optimizer finds the factor point at which every selected response is simultaneously as good as possible. Each response is mapped through a desirability function to d ∈ [0, 1], and the geometric mean D of those values is maximised over the factor box via multi-start Nelder–Mead.
Desirability function (Derringer-Suich): Three flavours: maximise (linear ramp between an unacceptable lower bound and a fully satisfactory upper bound), minimise (mirror image), and target-is-best (two-sided, peak exactly at the target). Shape parameters control how demanding the function becomes near the bounds.
Composite D: D = (∏ d_i^w_i)^(1/Σw_i). The moment any individual response hits d=0, D collapses to 0 — no single response is allowed to leave its acceptable range.
Models come from the Regression module (click "Save as model" there). The optimisation result itself is project-central, stored in state.optimizations[id].
Available in the following cycles
- DMAIC: Improve
- DMADV: Design
- 8D: D5 — Corrective Actions