Abstract
Task specialization is widely assumed to improve the performance of multi-robot systems, but it comes with hidden costs: coordination overhead, subtask dependencies, and increased design complexity. In this talk, I present a cost-benefit analysis of evolving specialized versus generalist robot controllers in a foraging scenario, showing that the evaluation budget required for specialists to outperform generalists decreases with increasing system size. I will then show how task specialization can also emerge spontaneously, driven by robot density and initial conditions, without being explicitly rewarded. Finally, I will outline ongoing work on whether self-organization in multi-robot systems can emerge from task-independent rewards.
About the speaker
Paolo Leopardi is a Ph.D. candidate in the Cyber-Physical Systems Group at the University of Konstanz and a member of the Center for the Advanced Study of Collective Behaviour. His research focuses on the emergence of division of labor, task partitioning, and task allocation in multi-robot systems, investigated mainly through evolutionary robotics approaches. He holds a Master's degree in Computer and Robotics Engineering from the University of Perugia, Italy. Since 2025, he has been a board member of the editorial team of the prestigious and worldwide recognized Sci-Phy seminar series.