AbstractsMathematics

Interest in adaptive and distributed systems for routing control in networks has led to the development of a new class of algorithms, which is inspired by the shortest path finding behaviours observed in biological ant colonies. This class utilizes ant-like agents, which autonomously traverse the network and collectively construct a distributed routing policy. Agent-based routing algorithms belonging to this class do not require a complete model of the network, and are able to adapt autonomously to network changes in dynamic and unpredictable environments. Previous studies of these algorithms have been carried out exclusively via the use of simulation-based models. In this thesis, we apply analytic modelling techniques to the study of agent-based routing algorithms. Our aim is to broaden the research in this field, as well as to gain a greater theoretical understanding of some fundamental properties of this class of algorithms.