Skip to Main Content
The paper provides a new paradigm to use Bayesian-net model to build a new multi-agent system (MAS). We use influenced diagrams as a modeling representation of agents, which is used to interact with them to predict their behavior. We provide a framework that an agent can use to learn the model of other agents in a MAS system based on their observed behavior. Since the correct model Is usually unknown with certainty, our agents maintain a number of possible models and assign the probability of being correct Our modification refines the parameters of the influenced diagram used to model the other agent's capabilities, preferences, or beliefs. The modified model is then allowed to compete with the other models and the probability assigned to it being correct can be reached based on how well It predicts the observed behaviors of the other agent.