Experience-based design of behaviors in videogames

Abstract

Artificial intelligence in games is usually used for creating player's opponents. Manual edition of intelligent behaviors for Non-Player Characters (NPC) of games is a cumbersome task that needs experienced designers. Amongst other activities, they design new behaviors in terms of perception and actuation over the environment. Behaviors typically use recurring patterns, so that experience and reuse are crucial aspects for behavior design. In this paper we present a behavior editor (eCo) using Case Based Reasoning to retrieve and reuse stored behaviors represented as hierarchical state machines. In this paper we focus on the application of different types of similarity assessment to retrieve the best behavior to reuse. eCo is configurable for different domains. We present our experience within a soccer simulation environment (SoccerBots) to design the behaviors of the automatic soccer players. © Springer-Verlag Berlin Heidelberg 2008.