The Biography of an Algorithm: Ant Colony Optimisation

Fri, 01/20/2017 - 14:00 to 15:30
ISG meeting room

Jérémy Grosman, PhD candidate

The walls of the Institut de Recherche Inter-Disciplinaire en Intelligence Artificielle (IRIDIA) at the Université Libre de Bruxelles (ULB), like now most research centres, are adorned of conferences’ posters. Strangely enough, third of these posters are about ants, just like third of the publications available in the coffee room and third of the conferences attended or organised. The engineers interested by ants are, roughly said, of two different sorts: those making optimisation algorithms and those working in swarm robotics. This paper focuses on “Ant Colony Optimisation”, a now a widespread algorithmic technique, which was developed around the mid-nineties, in what was about to become IRIDIA’s  “Optimisation Group”. What did these ants do to these engineers’ algorithms? In order to narrate the emergence of this algorithmic technique, we will be minutely following the successive transformations, which led from the ants themselves, to the entomologists’ natural ants and, ultimately, to the artificial ants of our engineers. While interesting in its own right, this historical account will prove chiefly instrumental in better characterizing the notions of algorithms and algorithmic practices.