Andrew Trusty, Santiago Ontañón, Ashwin Ram
We present a domain independent off-line adaptation technique called Stochastic Plan Optimization for finding and improving plans in real-time strategy games. Our method is based on ideas from genetic algorithms but we utilize a different representation for our plans and an alternate initialization procedure for our search process. The key to our technique is the use of expert plans to initialize our search in the most relevant parts of plan space. Our experiments validate this approach using our existing case based reasoning system Darmok in the real-time strategy game Wargus, a clone of Warcraft II.
Subjects: 1.8 Game Playing; 1.9 Genetic Algorithms
Submitted: Aug 12, 2008