er. Despite common

roots, AI soon dis-tinguished itself in its concern with autonomous prob-lem solving and its emphasis on symbol-ic, rather than numer-ic, information. Some of the earli-est AI research addressed approxi-mations and heuris-tics for complex tasks in decision-theoretic formula-tions (Simon 1955), and early work on expert systems for diagnosis (although before this term became popular) used Bayesian and decision-theoretic schemes (Gorry and Barnett 1968). However, many AI researchers soon lost interest in decision theory. This disenchantment arose, in part, from a perception that it was hopelessly intractable and inadequate for expressing the rich structure of human knowledge (Gorry 1973; Szolovits 1982). Although similar views are still widespread among AI researchers, there has been a recent resurgence of interest in the application of probability theory, decision theory, and deci-sion analysis to various problems in AI. In Decision analysis and knowledge-based expert systems share some common goals. Both technologies are designed to improve human decision making; they attempt to do this by formaliz-ing human expert knowledge so that it is amenable to mech-anized reasoning. However, the tech-nologies are based on rather different prin-ciples. Decision anal-ysis is the application of the principles of decision theory supplemented with insights from the psychology of judgment. Expert systems, at least as we use this term here, involve the application of various logical and computa-tional techniques of AI to the representation of human knowledge for automated infer-ence. AI and decision theory both emerged from research on systematic methods for problem solving and decision making that first blossomed in the 1940s. They even share a common progenitor, John von Neumann, who was a coauthor with Oscar Morgenstern of the best-known formulation of decision theory as well a key player in the developmen