Using preference based heuristics to control abductive reasoning

摘要

Explanations for symptoms can be selected using a variety of criteria. When numerical information is scarce, approaches which depend on partial orders could be of value. The problem of generating explanations for symptoms when only preference relationships between possible causes is considered. The lack of information leads to large number of possible solutions and control of the reasoning process is very important. A computational approach based on focusing the reasoning using a cost bounded ATMS is described.