CLUSTERING

Clustering is a method that aims to reduce the routing complexity by dividing the network in small subsystems called clusters, for each cluster a clusterhead is selected that works as the natural backbone of the network

flowchart TB
subgraph clster2
direction LR
E((node1))
F((node2))
G((node3))
H((clusterhead))
E & F & G --> H
end
subgraph cluster1
direction LR
A((node1))
B((node2))
C((node3))
D((clusterhead))
A & C & B --> D
end
D <--> H

there are several ways to elect a clusterhead

• LEECH a node make a local decision if it should serve as a clusterhead then nodes connect to the closest clusterhead, clusterhead is periodically reasigned to improve energy consumption
• HEED a node announce that it’s intention to become clusterhead and the other nodes votes the clustered with the minimal cost

CLUSTERING PROBLEMS

creating optimal clusters is an np-complete problem so forget about optimal clustering and in a Manet situation where nodes are highly mobile there is no certain that a cluster will not become sub-optimal in the long run.

Clusterhead plays a delicate role in a cluster, the role is suitable for less mobile nodes, with less energy restrictions

PREVIOUS NEXT