A mobile ad-hoc network (MANET) is a collection of mobile hosts that communicate with each other using wireless medium. An important problem in MANET is routing of data between these mobile hosts. Various different approaches were proposed to handle challenges of the routing problem, and recently, a new trend of approaches emerged to strengthen existing routing protocols by targeting more specific network properties. Routing in MANET, thus, is becoming an increasingly important research area with its applications that range from sensor networks to satellite ad hoc networks. Machine learning researchers, however, have not made major contributions to this area yet.

In this talk, we will present an overview about MANET routing protocols and major contributions made through application of swarm intelligence. We will discuss why MANET is an an interesting domain for the application of machine learning techniques and present potential ideas from temporal association and reinforcement learning that can be applied to mobile ad hoc routing.