MultiPoint Relay selection using GA

Abstract

Multi-Point Relay (MPR) mechanism is a selection algorithm to reduce the number of control packets in ad-hoc network routing protocols. This study presents a new MPR selection method using genetic algorithm (GA) while the MPR sets follow the maximal independent concept. Although original greedy algorithm works well in sparse networks, as MPR selection is an NP-hard problem, the proposed method can improve MPR selection phase in the presence of large number of nodes. The simulation shows that our method obtains up to 35% more independent sets when node density is high and incurs less memory usage, less power-consumption, and less retransmissions due to interferences.