DOE terms
Factorial, fractional, resolution, main effects, interactions, response surfaces.
A designed experiment changes several factors at once in a planned pattern, so their effects and interactions can be measured with far fewer runs than one-factor-at-a-time testing.
| Term | Meaning |
|---|---|
| Factor / level | An input you change (temperature) and the settings you test (180, 200 °C). |
| Response | The output you measure. |
| Full factorial | Every combination of k factors at 2 levels: 3 factors = 8 runs, 4 = 16, 5 = 32. |
| Fractional factorial | A planned fraction of the runs; some effects are confounded (aliased). |
| Resolution III / IV / V | III: main effects aliased with 2-factor interactions. IV: main effects clear. V: main effects and 2-factor interactions clear. |
| Main effect | Average response at the high level minus average at the low level. |
| Interaction | The effect of one factor depends on the level of another. |
| Center points | Runs at the middle level, to check for curvature and estimate error. |
| Randomization, replication, blocking | Run in random order; repeat runs; group runs to remove known nuisance effects. |
| Response surface (CCD, Box-Behnken) | Designs that fit curves, to find an optimum. |
