Design and analysis of a leader election algorithm for mobile ad hoc networks

Leader election is a very important problem, not only in wired networks, but in mobile, ad hoc networks as well. Existing solutions to leader election do not handle frequent topology changes and dynamic nature of mobile networks. We present a leader election algorithm that is highly adaptive to arbitrary (possibly concurrent) topological changes and is therefore well-suited for use in mobile ad hoc networks. The algorithm is based on finding an extrema and uses diffusing computations for this purpose. We show, using linear-time temporal logic, that the algorithm is "weakly" self-stabilizing and terminating. We also simulate the algorithm in a mobile ad hoc setting. Through our simulation study, we elaborate on several important issues that can significantly impact performance of such a protocol for mobile ad hoc networks such as choice of signaling, broadcast nature of wireless medium etc. Our simulation study shows that our algorithm is quite effective in that each node has a leader approximately 97-99% of the time in a variety of operating conditions.