Define a Study

When setting up a generative design study to explore design alternatives (outcomes), select the generation method and provide information required for the study.

Video: Create a Generative Design Study

For information about using the sample study types provided with the Revit installation, see Sample Study Types.

During the Create Study workflow, the Define Study dialog displays after you select the study type. Proceed as follows:

1. Select a generation method:
   • Optimize: Iterate to meet goals
   • Randomize: Use random values within ranges
   • Cross-product: Evenly distribute values within ranges
   • Like this: Explore variations of a preferred outcome

   The method determines the algorithm used to generate outcomes.

   Each method requires that you provide a different set of information. When you select a method, the Define Study dialog updates to display appropriate fields.

2. Select elements in the model or information to use in the design.

   If the Define Study dialog shows a section titled Select, you must select one or more elements in the model, select values from drop-down lists, or define a file path for an input node that references a file but has no file input data.

   Until you select these items, a warning symbol displays next to each item, and you cannot click the Generate button to start the study.

   Tip: If the warning symbol displays next to the Browse file path button, you can package the files required to run the Dynamo graph by going to Dynamo Generative Design Generative Design Tools and selecting the Include File Dependencies option. That file will be selected by default going forward. For more information about using Generative Design in Dynamo, see Create a Study Using Dynamo for Revit.

   To select elements from the model:

   1. Click the Select button.
   2. In the Revit model, select one or more elements, as appropriate for the study.

      Example: To generate room layouts for desks, select the room where the desks will be located.

      In some cases, you need to select multiple elements at once. For example, when creating a Maximize Window Views study, you select multiple windows or curtain elements. After doing so, click Finish on the Options Bar.

   3. Return to the Define Study dialog.

   After you select appropriate elements or values, a green checkmark displays in the dialog next to each Select item.

3. Specify goals, variables, constants, and constraints.

   The information required to run the study varies, depending on the chosen generation method.

   • Set goals. (Optimize method only) What are the goals of the design? Select the goals to use in this study, and choose Minimize or Maximize for each.
   • Set variables and constants. Which values can change to create different outcomes? Select the corresponding check boxes. The range of values is indicated next to each, as defined for the study type.

     For the Like This method, define the initial value for each variable. (Study values will vary by +/- 20% of the selected value.)

     For the Cross-Product method, specify the number of variations of each variable to use when generating outcomes.

     To define a constant, using the same value for all outcomes of the study, clear its check box, and specify the value to use.

   • Set constraints. (Optimize method only) What are the constraints or requirements of the design problem? Which values should be restricted within a certain range? Select those items, and specify the minimum and maximum values to use for the study. To specify a single value, enter the same number for the minimum and the maximum.
   • Generation settings. These settings determine how the study generates successive iterations of outcomes.

     In this topic

     • Population size
     • Generations
     • Number of outcomes
     • Seed
     • About genetic algorithms
   Tip: To save these settings as default values for this study type and the selected method, click (Study Options) to the right of the study name, and select Save Default Settings. Whenever you define another study using this study type, the default settings are used as the starting point. Change them as needed for each study and method.
4. Resolve issues.

   The Issues section lists any issues that must be resolved before you can start the study to generate outcomes.

   When all issues are resolved and required input is provided, the Generate button is enabled.

5. To start the study and generate outcomes, click Generate.

   This process may take a while. The amount of time required depends on the number of outcomes or generations to be created. If desired, you can minimize the window and resume work in the model, or use other applications. Check back occasionally to see if the generative design process is complete.

6. Explore outcomes.