Controlling Task Distribution in MONEE

The MONEE framework endows collective adaptive robotic systems with the ability to combine environmentand taskdriven selection pressures: it enables distributed online algorithms for learning behaviours that ensure both survival and accomplishment of user-defined tasks. This paper explores the trade-off that must be reached between these two (possibly contradictory) requirements, in the case where a foraging task is defined by the user. In particular, we study the impact of enforcing specialisation (i.e. the collective must acquire two mutually exclusive foraging skills) as well as the mechanism for tuning the level of specialisation in an on-line fashion. Results show that the actual behaviour of the collective system can be guided on request during the course of evolution in order to achieve a particular distribution of specialisations, albeit within a certain range of values.