Autonomie decision making based on bayesian networks and ontologies

In modern decision making systems, it is necessary to develop ways to optimize the decisions. To achieve this objective, new paradigms are necessary, as those linked to autonomic paradigm, which establishes, among other things, the self-configuration of the systems in order to improve their performances. On this basis, this paper proposes a system of autonomous decision, based on Bayesian networks and ontologies, to optimize and adapt a system to the characteristics of the context. The ontology provides the knowledge about the context to study, to configure the structure of the Bayesian network, while this latter is the stochastic reasoning mechanism used by the decision-making system. To verify the operation of our approach, it is used in an autonomic communication system. In the tests realized we could observe an improvement in the system performance.