Introducing Plausibility Measures in the Process of Revision of Beliefs Through Extended Logic Programs

. A new proposal for incorporating new evidence in an agent's knowledge base is presented in this paper. This proposal is based on a semi-qualitative notion of plausibility and on the abductive explanation for the newly observed fact. Plausibility is determined by specificity relations between the default rules in the agent's knowledge base and by the agent's contingential knowledge. We demonstrate how these specificity relations and the notion of plausibility degree can be represented by an extended logic program. 1 INTRODUCTION A rational agent should behave itself in situations where the available information concerning the environment and other agents is incomplete. To achieve this goal it is necessary that the agent foresees the real situation it is immersed in. However, these predictions may contradict new evidence, and should, therefore, be revised. And so, rational behavior must deal with the assimilation of new information. It is usual to distinguish two processes for incorpo..