ABSTRACT
Implementations of artificial intelligence (AI) often catalog a system's every state and event and use brute-force searches to find the event that is prede fined as optimal for a given configuration of states. The search proceeds as follows: If this configuration of states is manifest, then select this event; if that configuration of states is manifest, then select that event; and so on. This approach-called production-system modeling-is remarkably successful with fast computers and well-bounded systems.
