Résumé
This paper describes a bio-inspired method that enables the production of artificial multiagent organisms starting from a single agent. This method relies on two complementary computing approaches: (1) mimicking the functioning of segmen-tation genes and homeotic genes in the development of natural embryos and (2) using techniques from the evolutionary computing and the artificial embryogeny fields. The paper first details these mechanisms and then studies some possibilities of such a bio-inspired approach developing multiagent systems that evolve and self-organize in various flag patterns.