Modeling evolving behaviors in ant colonies

Abstract

Modeling societies of individuals is a challenging task increasingly attracting the interest of the Machine Learning community. In ant colonies, individuals have the same physical attitudes and, in principle, can assume any role the social environment requires. For he biologist understanding the complex dynamics ruling a colony is hard due to the difficulty of collecting and classifying long term ant activities. Here we present a novel approach for analyzing activity logs from an ant colony using an RFID environment. A semi-automated algorithm for segmenting traces and discovering the role played by each individual during the observation phase is described. A Structured Hidden Markov Model was used to build the model of single individual activities. Then, the global profile of the colony was traced during the emigration from one nest to another. The method provided significant information concerning the social dynamics of ant colonies.