
occasionally knocked off course by a random force, similar to a gust of
wind. Again, the result was a very believable, simple, decentralized simu-
lation of bird flock choreography.
Some of our own earliest experiments were derived from approxima-
tions of Reynolds/Heppner bird flocks; in fact, Heppner’s roost idea pro-
vided an inspiration that led us deep into the study of swarm intelligence
(Kennedy and Eberhart, 1995). In our very first experiments, populations
of “birds”flew in orderly flocking patterns. If birds could be programmed
to flock toward a roost, then how about having them look for something
like birdseed? It seems impossible that birds fl