Closed Systems Paradigm for Intelligent Systems

Abstract

Intelligent systems ought to be distinguished as a special type of system. While some adopt this view informally, in practice, systems engineering methods for intelligent systems are still centered around traditional systems engineering notions of engineering by aggregation of components. We posit that this traditional approach follows from holding a notion of open systems as the fundamental precept, and that engineering intelligent systems, in contrast, requires an approach that holds notions of closed systems as fundamental precepts. We take a systems theoretic approach to defining closed system phenomena and their relation to engineering intelligence. We propose the concept of variety; particularly the law of requisite variety to enable closed view in engineering. We discuss how open and closed view approaches to engineering intelligent systems address variety differently, as well as the implications of this difference on engineering practice.