A.I. Techniques for Modelling Anger in Emotional Agents

Abstract

The research presented here, attempts to review a range of techniques commonly categorized under the umbrella of artificial intelligence (A.I.) that could be applied when developing agents with emotions in a range of applications. The paper focuses on anger (and its related emotions), an emotion strongly linked with aggression which of course forms the basis of many computer games where killing or attacking other players or in-game agents is often central to the game’s purpose. The paper begins with a psychology focused review of anger and its related emotions, before presenting techniques to encode some of these elements using Finite State Machines and Fuzzy Logic.