Computer Science

From Birds to Boards

See allHide authors and affiliations

Science  20 Oct 2006:
Vol. 314, Issue 5798, pp. 389
DOI: 10.1126/science.314.5798.389b

One way to simulate a complex phenomenon such as the flocking of birds is to build a detailed computer model that incorporates the motion of all individuals as well as rules for their interaction. Many aspects of the underlying mechanisms must be incorporated into the program from the start, and such a system could be susceptible to wildly nonlinear outcomes. An alternative is to build a large number of autonomous entities having no central controller, and then allow the collective behavior to emerge over time through trial and error. As Liu and Tsui explain, such computing models have been labeled “nature-inspired” because they are truly analogous to the way birds flock or ants in a colony engage in purposeful collective behavior. The basic entities in such a model (which could be either software modules or actual hardware robots with internal software) are self-organized, most strongly influenced by local interactions, and capable of adapting their behavior in response to a changing environment. These emergent adaptive computing systems may find applications in studies of such complex collections of interacting entities as the human immune system, economic markets, or ecological communities. — DV

Commun. ACM 49, 59 (2006).

Navigate This Article