Discrete Event Abstraction: An Emerging Paradigm For Modeling Complex Adaptive Systems

Computer modeling and simulation is recognized by John Holland and many others as the central tool with which to experiment on complex adaptive systems (CAS). Less well recognized is that in the last thirty years, advances in the theory of modeling and simulation have crystallized a new class of models suitable for the computational requirements of CAS studies. This article discusses the abstractions underlying the DEVS formalism, a system theoretic characterization of discrete event simulations, that has been widely adopted in recent years. Abstraction of events and time intervals from a continuous data stream is shown to carry information that can be efficiently employed, not only in simulation, but also in accounting for the real world constraints that shape the information processes within CAS. Indeed, an important paradigm is emerging in which discrete event abstraction is recognized as fundamental to modeling CAS phenomena at various levels of organization. Discrete event models of neurons, neural processing architectures, and "fast frugal" bounded rational decision making and shortest path solvers are discussed as examples. Such models capture ideas that are coming from various disparate directions and offer evidence that a new modeling and simulation paradigm is emerging.