Backbone formation in Wireless Ad-hoc Networks using Distributed Q-learning

Unlike a wired network, there is no base station or network infrastructure in a wireless ad-hoc network. In order to make routing and flooding efficient in such networks, Connected Dominating Set (CDS) as a virtual backbone has been extensively studied. Clustering the Ad-hoc networks by finding the Minimum Connected Dominating Set (MCDS) is a promising approach to construct an efficient virtual backbone in wireless Ad-hoc Networks. Since in such networks, there is neither a fixed infrastructure nor a central administration, the centralized algorithms are not feasible in such environments. Therefore, in this paper, a distributed algorithm based on distributed Q-learning, called DQ-CDC, is proposed for finding the MCDS in Wireless Ad-hoc networks. The simulation results show the superiority of the proposed algorithms over the existing CDS-based backbone formation algorithms.