A Systemic Metaphor of Multi-agent Coordination in Living Systems

Living systems provide a good metaphor for the coordination of autonomous agents managing tasks according to their role in the system. The study of such systems from the point of view of General Systems Theory would benefit from computational experimentation. This paper proposes an agent-based methodology for system modeling as well as computational tools to develop and test system models. Biological problems are general enough to provide a metaphor by which one can model several problems in coordinating multiagent behavior. We show that systems theory and the study of living systems provide insight in the various problems of multiagent coordination, and that this insight produces computational models where we can express complex coordination patterns involving highly-specialized agents, particularly those cases where the issue of sharing limited resources is crucial.