A cluster based leader election algorithm for MANETs

Various Ideas on leader election process has evolved and most of them rely on techniques such as extrema finding or centrality finding methods. These methods are very reliable and highly efficient in a static network, but when discussion reaches ad-hoc networks robustness is questioned and current system doesn't answer to this dilemma. So taking in to account this highly dynamic nature of mobile ad-hoc networks current scenarios of leader election must be modified. This paper describes a novel method to cope up the dynamic nature of ad- hoc networks by proposing some changes to the current leader election process. The main idea is to cluster the networks and to use an `avoidance strategy' for the clusters up to a particular threshold, and thus neglecting the dynamicity of the sub networks during the leader election process without affecting the performance.