About agents that reason by case (Preliminary report)

This paper is concerned with intelligent agents that are able to perform nonmonotonic reasoning in the presence of various sources of information. More precisely, the focus is on default logic agents. It is well-known that default logic suffers from a lack of reasoning by case ability. Such a problem is even more important in the context of intelligent agents that should be able to draw the most valuable conclusions from several information sources. In the paper, an original approach to address this default logic drawback is investigated. Interestingly, it allows several levels of reasoning by case to be differentiated.