When creating a Generative Design study, consider the criteria you will use to generate design alternatives, including goals, constraints, variables, and constants. Understand the outputs that result for each outcome.
In this topic
The design criteria for a study depends on the study type (how the Dynamo graph is written), as well as the generation method that you select. For more information about design criteria, see Generative Design Primer: Anatomy of a Good Generative Design Process.
The following descriptions use a sample Three Box Massing study to illustrate design criteria.
Inputs
Inputs are data to be used when generating outcomes. The data may be provided by you, or it may be defined by the study type (in the Dynamo graph).
Inputs can include the following:
- variables
- constants
- other supported input types (file paths, unit inputs, etc.)
Variable
A variable is a value that can change when generating outcomes.
For the Three Box Massing study, for example, you can specify that the following are variables:
- box height
- box position
When defining the study, you indicate which values are to be variables for the study.
To get a quick understanding of the range of variable values used to generate outcomes for a study, look at its parallel coordinates chart on the Explore Outcomes dialog.
Constant
A constant is a value that is fixed and does not change. A constant remains the same for all outcomes.
These values may be programmed as part of the Dynamo graph so you can't change them.
As an alternative, you can specify a constant value when creating a study. The same value is used for all outcomes of the study.
Goal (Optimize method only)
A goal is an objective that you want the design to achieve.
Goals are used by the Optimize generation method. This method improves succeeding generations of design alternatives based on prior results.
For example, suppose you are creating a Three Box Massing study to generate alternatives for three adjacent buildings (as simple masses). This study type has the following goals:
- Maximize or minimize floor area
- Maximize or minimize surface area
These goals may conflict: You may want to maximize floor area to increase rental space while minimizing surface area to reduce construction costs. The Optimize method works to find a solution that maximizes the floor area while also making sure that the surface area remains as small as possible.
See also Generative Design Primer: Defining Goals.
Constraint (Optimize method only)
A constraint is a condition that a design alternative must satisfy.
Constraints confine design alternatives to values within a specified range.
The following are sample constraints for the Three Box Massing study:
- The floor area must be at least 60,000 square feet.
- The maximum surface area is 80,000 square feet.
When a study is unconstrained, it may return unrealistic, impractical results, such as a surface area that is too large for the project specifications. Constraints ensure that the study generates design alternatives that are reasonable for a particular situation.
Generation Settings
When using the Optimize, Randomize, or Like This method, you can change generation settings. These settings can include population size, generations, number of solutions, and seed. Learn more about generation settings.
Outputs
For each outcome, certain outputs are calculated during the generative design process. These outputs reflect the specific values measured for a particular outcome.
For example, a single outcome of the Three Box Massing study includes the following outputs:
- floor area
- surface area
You might use some of these outputs as inputs to another study using the Like This generation method, refining the design criteria with each iteration.
Outcome
An outcome is a design alternative generated by Generative Design as part of a study. See the outcomes in the Explore Outcomes dialog.
