Failing Queries in Distributed Autonomous Information System

There are two basic cases when Query Answering System (QAS) for a Distributed Autonomous Information System (DAIS) may give no answer to a submitted query. Let us assume that q is that query which is submitted to an information system S representing one of the sites in DAIS. Systems in DAIS can be incomplete, have hierarchical attributes, and we also assume that there are no objects in S which descriptions are matching q. In such a case, QAS will fail and return the empty set of objects. Alternatively, it may relax query q as it was proposed in [7], [8], [2]. It means that q is replaced either automatically or with a help from user by a new more general query. Clearly, the ultimate goal is to find a generalization of q which is possibly the smallest. Smaller generalizations of queries always guarantee higher confidence in objects returned by QAS. Such QAS is called cooperative. We may also encounter failing query problem when some of the attributes listed in q are outside the domain of S. We call them foreign for S. In such a case, we extract definitions of foreign attributes for S at other sites in DAIS and next used them in QAS to solve q. However, to do that successfully, we have to assume that both systems agree on the ontology of their common attributes [14], [15], [16]. Such definitions are used to identify which objects in S may satisfy that query. The corresponding QAS is called collaborative. This paper shows that cooperation can be used as a refinement tool for the collaboration strategy dealing with failing query problem as presented in [14], [15].