December 2010 | Volume 13 Issue 4 36 FELLOWS' INSIGHT ized to form multicellular animals, and an animal may be an agent in a formation of a particular herd of animals.
Systems and the Relationship to HumansA traditional view of systems considers the system to be a technological or physical system, with humans as operators and observers being considered outside the system, whereas most work with CASs includes the notion of human activity systems within the CAS itself. A system may be complex and intricately patterned. In contrast, the observer may be facile (able to navigate such complexity) or may believe the system ought to be orderly and mechanically predictable, and falsely attribute "unpredictability" to the system. This gives rise to the notion that complexity lies not within the system being observed nor within the observer but in the nature of the relationship between the observer and the observed.
Complex Systems Disciplines and What Makes a System ComplexMost studies of complex systems (such as Axelrod and Cohen 1999) are multidisciplinary, which runs counter to a general scientific trend toward increasing fragmentation and specialization in most traditional disciplines. It is not the large number of parts in a system that makes the system complex; it is the way that the parts interact. A product may consist of abundant parts, but if these parts interact only in a known, designed, and structured fashion, then the system is not complex, although it may be big. Some researchers have divided the problems of large systems into two types: "complicated" for systems that have many elements though their relationships are only moderately tangled, and "complex" for systems whose relationships are hard to tease apart and indeed may contribute most of the value to the system.
Relevance to the Practice of Systems EngineeringThese notions suggest that the practice of systems engineering needs to include more knowledge from the field of complex systems than has survived the last two decades of process-orientation. The theory behind what we do must be made explicit. Our heuristics must also be resolved against theory. Our cause-effect approach must broaden and also attend to critical-mass and tipping-point concepts, and our modeling must broaden from functions and flows to capabilities and effects. Furthermore, predicates are not sufficient in CAS efforts, including design. Designers must admit to the presence of concurrency and simultaneity.
Issue 5: "What Distinguishes Complex Adaptive Systems From Other Kinds of Systems?"Andrew P. Sage, andrew.sage@incose.org; Jack Ring, jack.ring@incose.org; and Sarah Sheard, sarah.sheard@incose.org
Central Ideas and Concepts in Complex Adaptive SystemsC omplex adaptive systems (CASs) are currently being studied in many places, including universities in many different departments. As a result, what are considered the central ideas of this field vary considerably from one research group to another, making it difficult to capture specific, agreed-upon definitions. Nevertheless, som...