Routing in ad-hoc networks using minimum connected dominating sets

Deterministic Dominating Set Construction in Networks with Bounded Degree

“…Kuhn et al [19] showed that any distributed approximation algorithm for the dominating set problem with a polylogarithmic approximation ratio requires at least Ω( log n/ log log n) communication rounds. Along with that, the existing distributed deterministic algorithms incur a linear (in number of nodes) running time [7,23,29]. This worst-case upper bound remains valid even when graphs of interest are restricted to the bounded degree case, like the ones described above.…”

“…This worst-case upper bound remains valid even when graphs of interest are restricted to the bounded degree case, like the ones described above. The deterministic approximation algorithms [7,23,29] are based on the centralized algorithm of Guha and Khuller [13], which in turn is based on a greedy heuristic for the related set-cover problem [5]. Following the heuristic, these algorithms start with an empty dominating set and proceed as following.…”

“…The decision whether to join the dominating set in the above iterative process is taken based on the lexicographic order of the pair span, ID [7,23,29]. The use of IDs to break ties leads to long dependency chains, where a node cannot join the set because of another node having higher ID.…”

Routing Transient Traffic in Mobile Ad Hoc Networks

“…Another set of integrated approaches are the spinebased algorithms from R. Sivakumar et al [27,29]. Compared to HSR and HierLS, the spine approach is simpler, because it uses only two layers in the routing hierarchy.…”

“…One is to reduce the backbone complexity by building a small connected graph in the network. Building a backbone that has good approximation ratio to the ideal Minimum Connected Dominating Set (MCDS) has been the research focus of ad hoc clustering in recent years [22][23][24][25][26][27][28]. Another method is to reduce the protocol overhead forwarded on the backbone.…”