CommonKADS and cooperating knowledge based systems

In this paper we shall deal with the knowledge acquisition for cooperating knowledge based systems using a preliminary version of the CommonKADS methodology with minor extensions. The main result of our study is the description of the task, inference and domain knowledge of a general supervisor agent which is responsible for communication and cooperation in our CKBS. According to the CommonKADS philosophy task and inference knowledge of such a general supervisor agent are reusable indepently of the underlying CKBS shell. We will show the applicability of CommonKADS for CKBS Knowledge Acquisition in a macroscopic sense. Particulariliy on the domain layer of the expertise model there is a need to provide a model for reasoning with di erent models of uncertainty. During the result recomposition stage the supervisor agent has furthermore to deal with inconsistent knowledge due to the di erent viewpoints that agents in an overlapping application area might have, e.g. di erent doctors.