On the Cost of Evolving Task Specialization in Multi-Robot Systems
Authors: Paolo Leopardi, Heiko Hamann, Jonas Kuckling, Tanja Katharina Kaiser
Organizations: Centre for the Advanced Study of Collective Behaviour, University of Konstanz, Konstanz, Germany · Department of Computer and Information Science, University of Konstanz, Konstanz, Germany · Department of Computer Science and Artificial Intelligence, University of Technology Nuremberg, Germany
Task specialization can lead to simpler robot behaviors and higher efficiency in multi-robot systems. Previous works have shown the emergence of task specialization during evolutionary optimization, focusing on feasibility rather than costs. In this study, we take first steps toward a cost-benefit analysis of task specialization in robot swarms using a foraging scenario. We evolve artificial neural networks as generalist behaviors for the entire task and as task-specialist behaviors for subtasks within a limited evaluation budget. We show that generalist behaviors can be successfully optimized while the evolved task-specialist controllers fail to cooperate efficiently, resulting in worse performance than the generalists. Consequently, task specialization does not necessarily improve efficiency when optimization budget is limited.
Figures & tables
Figure 1: Overview of the foraging task: Robots have to transport objects (squares) from the source across a slope to the nest. The task can either be executed by each robot individually (generalist) or shared between robots (dropper and collector).
Figure 2: Best fitness (Eqs. 3 , 4 ) for the generalists, droppers, and collectors over 100 generations. Dashed lines: different independent evolutionary runs; black line: mean.
Figure 3: Performance of groups of 2 robots. Letters indicate the subtask performed (generalist G, dropper D, collector C), numbers indicate the evolutionary run. Each letter-number pair is associated with a robot.
Centre for the Advanced Study of Collective Behaviour & Department of Computer and Information Science University of Konstanz Konstanz, Germany · Department of Computer Science and Artificial Intelligence University of Technology Nuremberg Nuremberg, Germany