Belief-Desire-Intention Deliberation in Artificial Soccer

Many different architectures have been proposed for the design of autonomous agents. In this article, the application of the belief-desire-intention architecture to the artificial soccer domain is described. We show how it supports efficient deliberation in a highly dynamic environment. The mapping from input to output can be done in simple ways (for example, strictly reactive) or in more sophisticated ways, for example, with models that are inspired by human decision processes. We found that mental notions such as capabilities-skills, belief, goals-desires, and intentions-plans are useful pictures to make agent programming transparent.