The Origin of Rule-Based Systems in AI

Since production systems (PS’s) were first proposed by Post (1943) general computational mechanism, the methodology has seen a great deal of development and has been applied to a diverse collection of problems. Despite the wide scope of goals and perspectives demonstrated by the various systems, there appear to be many recurrent themes. We present an analysis and overview of those themes, as well as a conceptual framework by which many of the seemingly disparate efforts can be viewed, both in relation to each other and to other methodologies. Accordingly, we use the term production system in a broad sense and show how most systems that have used the term can be fit into the framework. The comparison to other methodologies is intended to provide a view of PS characteristics in a broader context, with primary reference to procedurally based techniques, but also with reference to more recent developments in programming and the organization of data and knowledge bases. This chapter begins by offering a review of the essential structure and function of a PS, presenting a picture of a "pure" PS to provide a basis for subsequent elaborations. Current views of PS’s fall into two distinct classes, and we shall demonstrate that this dichotomy may explain much of the existing variation in goals and methods. This is followed by some speculations on the nature of appropriate and inappropriate problem domains for PS’s i.e., what is it about a problem that makes the PS methodology appropriate, and how do these factors arise out of the system’s basic structure and function? Next, we review characteristics common to all systems, explaining how they contribute to the basic character and noting their