Regional consecutive leader election in mobile ad-hoc networks

Abstract

In this paper we introduce the regional consecutive leader election (RCLE) problem, which extends the classic leader election problem to the continuously-changing environment of mobile ad-hoc networks. We assume that mobile nodes, including the currently elected leader, can fail by crashing, and might enter or exit the region of interest at any time. We require the existence of certain paths that ensures a bound on the time for propagation of information within the region. We present and prove correct an algorithm that solves RCLE for a fixed region in 2 or 3-dimensional space. Our algorithm does not rely on the knowledge of the total number of nodes in the system nor on a common startup time. In the second part of the paper, we introduce a condition on mobility that is sufficient to ensure the existence of the paths required by our RCLE algorithm.