Creation of desirable complexity: Strategies for designing selforganized systems

Abstract

Thirty years ago, Barry Malzberg described what may have been the first vision of aself-organized application: "surely ant and man could coexist peacefully ... we might even be able to voyage to the stars together, the ants developing a communications network that would implement our vast technological resources" [53, p. 118]. Although we have indeed developed ant-based routing for telecommunications [17, 32, 63], more generally progress has been disappointing. To date, self-organization research has focused mostly on dissecting and understanding these complex systems, elucidating the links between the micro and macro levels, and identifying their general characteristics and ingredients. We have lingered in the proof-of-concept stage, demonstrating that simple agents with local interactions and feedbacks can lead to complex, adaptive group-level properties. However, despite various authors waxing lyrical about the enormous potential for problem-solving self-organized systems [15, 33, 46, 47], there are surprisingly few practical applications in use. © 2006 Springer.