Maastricht University, The NetherlandsPieter Spronck
Real-time strategy (RTS) games are highly complex. They have extremely large decision spaces that cover military and economic research and development, civilization building, exploration, and sometimes diplomacy. They are also highly chaotic, and the outcomes of actions are often uncertain. This makes designing the AI for an RTS game extremely challenging and time consuming.
In this article, we discuss how a genetic algorithm can be used to help AI developers create effective RTS AI by automatically generating strong strategies. We concentrate on the representation of a strategy in the form ...