2.42.6.1 The Optimize Response Dialog Box

Supporting Information

To open the Optimize Response dialog

• Activate a report sheet generated from the Analyze Design, Analyze Variability or Analyze Binary Response‎
• Select Statistics: Quality Improvement: Design of Experiment from Origin menu
• Select the Optimize Response icon

  

The feature is available only for Definitive Screening Design Factorial designs Response surface designs Mixture Design

Topics for Further Reading: Introduction: Optimize Response Overview of Optimize Response Results

Goal

Setup Table

Response	The response(s) selected from your stored model; click the row to edit its settings
Goal	Minimize Target Maximize
Target	The value to hit - required and enabled only when Goal = Target
Lower Limit	Boundary below which the response is unacceptable
Upper Limit	Boundary above which the response is unacceptable
Weight	Shape of the desirability curve within the limits
Importance	Relative priority of this response vs. others (1 = equal) when multiple responses are optimized

Buttons

View Model	Opens the stored model equation for the selected response so you can verify the fitted terms before optimizing
Set Lower/Upper Limits	Automatically populates limits from the actual response data range

Settings

The Settings tab configures how the optimization algorithm searches for the best factor settings.

For Mixture Design, only the Iterations setting is available - the Constraints Table and Initial Value options do not apply to mixture designs. All other design types have the full set of options described below. Iterations:This sets the maximum number of iterations the optimization algorithm performs when searching for the best component proportions.

Constraints Table

Factor	Factor name from the stored model
Constraint	Dropdown to choose how the factor participates in the search No constraints Factor is free to vary over its full experimental range Hold at value Factor is fixed; enter the value in the Hold Value column. Note: The behavior for categorical factors depends on the design type Response Surface / Definitive Screening Design: A categorical factor can be held at multiple levels simultaneously (e.g., Reagent held at both TypeA and TypeB). The search evaluates each held level. To hold multiple levels, select them from the dropdown - multiple selection is supported. Factorial: A categorical factor can be held at only one level (e.g., Reagent held at TypeA or TypeB, but not both). Constrained to range Enter Lower Bound / Upper Bound to restrict the search to a sub-region of the design space (e.g., for safety, equipment, or cost reasons)(Numeric factors only)
Hold Value	The fixed level for a factor that is not optimized
Lower Bound	Lowest numeric value the optimizer may use (numeric factors only)
Upper Bound	Highest numeric value the optimizer may use (numeric factors only)

Initial Value

The toggle enables custom starting points for the optimization algorithm.

• On: Enter an Initial Value for each free numeric factor.
• Off: Origin automatically selects starting values across the factor space

Output

The tab controls which elements appear in the results. See Overview of Optimize Response Results for guidance on interpreting the results.

95% Confidence Limits	Adds the LCL/UCL columns to the Prediction table
95% Prediction Limits	Adds the LPL/UPL columns to the Prediction table - the interval for individual confirmation runs
Optimization Plot	Generates the interactive optimization graph with factor sliders and fitted response curves

Note: For Analyze Binary Response and Analyze Variability, the Prediction table does not include 95% LPL/UPL prediction limits - the 95% Prediction Limits option does not apply to these analysis types.