Adaptive Approximate Community Detection Algorithm for Bubble Rap Routing Protocol

Abstract

A social graph is a network that shows the representation of how the nodes are connected. The classification of the network is done into different communities for data transmission. In the Bubble Rap routing protocol, the local and global centrality of a node is considered for message transmission. The global centrality tells how much the node is important in the entire network while the local centrality tells how much a node is important in its community. In Bubble Rap routing algorithm communities are formed with the help of K-Clique algorithm. The K-Clique algorithm is mainly designed for binary static graphs whereas Delay-Tolerant Network is highly dynamic in nature as the connections among the nodes keep on changing with time. In an Adaptive Approximate community detection algorithm, the information of the current network and previous community structure is followed to deduce the current structure of the community for the network. The community is decided using node degree and a frequency-duration utility. In this paper, we have implemented the algorithm for bubble rap protocol, and depending on the results of simulation it can be seen that performance-wise this algorithm is a better thank-clique community detection algorithm.