The complexity of Multi Robot Sistems in general, and of Swarm Robotic Systems in particular, makes any comparison among learning strategies a very hard task. Our work intends to propose a new analysis methodology that allows different learning strategies to be compared on a common test field. We also define specific parameters that could be used to evaluate the performances of each new algorithm. We show the effectiveness of our methodology by comparing three different learning algorithms.
Swarm Robotics, learning algorithms, formal analysis
Marcello Cirillo. (2005)
A Formal Approach to the Comparison of Control Algorithms for Swarm Robotics. Master Thesis. Politecnico di Milano.