Resumen de An architecture for multireasoning knowledge-based systems

Juan Pedro Caraça-Valente Hernández , Daniel Borrajo Millán

• Although multireasoning systems have been a research area of interest during the last few years, most of the architectures that have been designed present several characteristics in common like predetermined multireasoning strategies, the impossibility to adapt the reasoning strategy to each problem and lack of flexibility in the multireasoning structure. In this paper we propose an architecture for Multireasoning Knowledge-Based Systems and the corresponding multireasoning model that can cope with various different inference techniques. In addition, the model also defines the reasoning level of the KBS. This model provides an explicit multireasoning representation structure that enables an easy modification of the reasoning modules available to the KBS. Its main characteristic is that the multireasoning strategy is absolutely data-driven: the decision of which reasoning method to use is based on the current state of the knowledge base and the history of the current session.