Behavior Anticipation Based on Beliefs, Desires and Emotions

Most of the existing models of intelligent software agents fail to consider an important aspect of human behavior, namely the impact of emotions on processes such as motivation, decision-making, planning, learning, and anticipation. The paper presents an emotional reasoning model of artificial agents, called Belief-Desire-Emotion( BDE). The model is built upon the influential Belief-Desire-Intention agent architecture and follows the Ortony-Clore-Collin's cognitive appraisal theory of human beings. We describe the different stages of the emotion generation process and emphasize how this process influence theoretical reasoning, such as belief and desire revision, and practical reasoning, such as means-end analysis. Additionally, we propose a set of basic emotions, and we exemplify how they are generated and the way they influence the behavior of our BDE agents.