In the current issue of "Science News" (November 11, 2000, Vol. 158, No. 20,
pp.314-316) is an article by Ivars Peterson describing work by Laurent
Keller and coworkers at the University of Lausanne, Switzerland, in which
small (2 inches on a side) motile robots were used to reproduce the foraging
behavior of social insects.

The little robots had a "nest", a means of sensing the level of "food"
stored in the nest, and a switch which triggered foraging behavior when the
"food" level fell to a set point (different for different individuals).
When a robot was activated by the closing of this switch, it left the nest
and roamed the area (9 square meters) searching for tokens representing
"food". When it found "food" it transported the token back to the "nest".
In some of the robotic experiments, "... one robot could enlist another if
it happened to identify a resource-rich area." [method of communication was
not described.]

The point of the work was to investigate ways to develop robust
self-organizing systems for finding alternative traffic routes over
congested telephone lines and novel algorithms for governing how robots
operating completely independently would work together.

According to Peterson, the effort was inspired by observations of the
activities of ants.  There is apparently a goodly amount of work of this
sort being done, which has been reported at three recent conferences (two in
Paris, one in Brussels) dedicated to this general subject-area.

The whole idea seems to fit with the opinion held by many that social
insects are simple organic robots, pre-programmed and unconscious, that
respond to stimuli (and combinations of stimuli) in set ways.  There is
nothing resembling planning, decision-making, or control of one individual
by others in any of the complex behaviors of a colony; just the meshing of
individual responses to environmental cues.

I wish the article had described the means of recruitment of additional
foragers.

Walter Weller