Résumé
Numerous studies have highlighted the impact of interpersonal relationships in the dynamics of innovation adoption and diffusion. It is therefore natural that agent-based simulation is a commonly chosen approach for studying these dynamics, due to its ability to represent individual decisions and interactions between individuals. Some of these studies have focused specifically on the impact of social influences on the construction of an opinion on innovation. However, these works use a very abstract and simplified representation of the social influence process, which greatly limits the type of analysis that can be done, particularly on the diffusion of specific messages such as arguments defending or criticizing innovation. In this paper, we propose a model, MIDAO, based on the theory of planned behavior and formal argumentation, which aims to go further in this field by explicitly representing the arguments exchanged between agents. We have carried out a series of experiments demonstrating the importance of the messages introduced about the innovation, and the way in which they are disseminated, on innovation adoption.