Objects, Rules, and Process Control

Abstract

Artificial Intelligence (AI) suggests the paradox of the solution of problems not subject to any systematic analysis. In Process Control, this impression is attractive - we need solutions. It is also unacceptable - we need formal confidence in our designs. This paper adopts three bases to overcome the paradox: *The formal analysis of such AI structures as Rules and Objects; *The analysis of higher level control system issues that might be fruitfully modeled by AI structures; *The discussion of computer-aided methods for analyzing and validating Process Control AI systems. Specifically, this paper examines Objects and Rules as examples of Self-Marshaling Functions, and their utility for modeling, documenting, and implementing control systems.