- Thinking about using Processing for simple web-based example demos of partial solutions (i.e. individual steering behaviors) and successful crowd phenotype demos
- And using Open Frameworks for the final EvoDesign application (web-interaction will need to be sacrificed for speed)
- I'll need to do a performance test to see if this is a good choice...
- Each subgroup of individuals in a crowd has their own genotype + finite state diagram
- States in the finite state diagram represent behaviors
- Each state has a vector of weights (gene) corresponding to behaviors in Reynolds's taxonomy
- The collection of state genes for one finite state diagram comprises a genotype.
- Transitions can be triggered by events (collision or proximity) or by time (frame == 24)
- Topography + transitions are directly specified by the user
- How should groups be initialized and/or changed? Perhaps initialized by the designer with the option for the computer to change memberships during the time seqence?
Demos
