Motivation and Evolutionary Pedagogical Agents

Motivating students is a major issue for current Intelligent Tutoring Systems. Modern theories of motivation (such as the Self Determination Theory) have shown the positive motivational impact of autonomysupport. Following this idea, we have proposed in a previous work an autonomy-supportive motivational design for eLearning systems. Later, the importance to take learner’s culture into account appeared obvious to us if we wanted to correctly enhance/maintain the learner’s motivation. Both these findings resulted in an ITS called MOCAS (Motivationally and Culturally Aware System). MOCAS is basically composed of a virtual world in which several pedagogical agents with different roles, behaviors and knowledge cooperate to provide a motivational (i.e. autonomy-supportive) and culturally adapted teaching to learners. In the real world, pedagogical behaviors frequently determine whether a preceptor will be well accepted/respected or not by his/her pupils (and in many cases, this rating will be culturally dependant). But whatever preceptors are rated, they can not change from day to day what they truly are and how they behave. In Intelligent Tutoring Systems, pedagogical agents could have this ability to dynamically evolve. In this paper we define how to genetically adapt the crowd of pedagogical agents that is inside our MOCAS. This process allows the production of generations of pedagogical agents whose behaviors are more and more fitting the learners’ motivational and cultural needs.