Applying Behavior Trees to StarCraft AI

Student thesis: Master Thesis and HD Thesis

  • Kenneth Sejrsgaard-Jacobsen
  • Søren Larsen
  • Torkil Olsen
  • Long Huy Phan
  • Jonas Øgendahl Groth
1. term, Computer Science, Master (Master Programme)
We investigate the area of game AI with focus
on the video game genre of Real-Time Strategy
using StarCraft as test platform. We proceed
to discuss different AI methods that have
an application in RTS games. We choose to
focus on behavior trees and present our implementation
of a behavior tree framework, for
implementing behavior trees, and an editor for
designing them. We test the framework by implementing
a set of behaviors described in the
report and finally conclude on the work done
by proposing new possibilities for further development.
Publication date22 Dec 2010
Number of pages70
ID: 42802558