Net-based computational models of knowledge-processing systems

Net representation, which provides a uniform semantics to a wide class of computational problems, such as numeric functions, logical systems, and models of dynamic and distributed systems, is reviewed. Using a uniform formalism of colored propositional nets, the methodology of net models is extended to represent Horn clauses, non-Horn clauses, and expert systems. It is shown that, using this formalism, tools can be built for expanding and partitioning knowledge-bases, and testing knowledge-base properties such as consistency, redundancy, and deadlocks.<>