Inference in Inheritance Networks using Propositional Logic and Constraint Networks Techniques

It's not my default: The complexity of membership problems in restricted propositional default logics. In AAAI-90: Proceedings of the 8th national conference on artiicial intelligence, Figure 4: Interaction graph for the theory presented in example 4.6 when the constraint graph is a tree, satissability can be determined in linear time. In Ben-Eliyahu and Dechter, 1991a] we have demonstrated the potential of this approach using one speciic technique called Tree-Clustering Dechter and Pearl, 1989], customized for solving propositional satissability, and emphasized its effectiveness for maintaining a default database. We have also characterized the tractability of the default theories as a function of the induced width 4 5 , w , of their interaction graph. We next generalize those results for network theories: The interaction graph of a network theory is an undi-rected graph where each literal in the theory is associated with a node, and for each p and for every = : =p in D, every q 2 and every r such that r 2 , there are arcs connecting all of them into one clique with p. Also, for each p?!q in W, there is an arc between p and q.