Some years ago I have worked (together with Owen Holland) on a project applying the principle of trail laying by ants to routing and load balancing in communications networks, which is known to be a complex problem. 

Evolution came up with a very elegant solution for routing in nature: pheromone trail laying by social insects.  It turned out that the same basic principle can be applied to route traffic in a network so that the load on network nodes gets nicely balanced, avoiding bottleneck links.

Various researchers are currently looking into this, while I have moved into other areas.  This page contains links to my publications, and the work of others.

Publications:

Schoonderwoerd, R., Holland, O.E., Bruten, J.L., & Rothkrantz, L.J.M. (1996) Ant-based load balancing in telecommunications networks.  Adaptive Behavior, vol 5, issue 2, p.169-207.  An earlier version of this paper is available as HP Technical Report HPL-96-76.

Schoonderwoerd, R., Holland, O.E., Bruten, J.L. (1997) Ant-like agents for load balancing in telecommunications networks.  The First International Conference on Autonomous Agents. ACM Press. (postscript version). 

Schoonderwoerd, R. & Holland, O.E.  (1999) Minimal agents for communications networks routing: The social insect paradigm.  In Hayzelden, Bigham (ed.) Software Agents for Future Communication Systems.  Springer Verlag.  Abstract here.

Other links:

My M.Sc. Thesis (1996): Collective Intelligence for Network Control. Unpublished Ir.-thesis.
Delft University of Technology, Faculty of Technical Informatics, May '96. (1769k postscript) or
(576k compressed postscript).

Mark Ward. There is an ant in my phone. In New Scientist, 24 January 1998. (web-version here).

Marco Dorigo's Ant Colony Optimisation page.

ENST Bretagne SWARM project.