Abstract
Ad-hoc networks can operate even when there is a scarcity of network infrastructure. This research examines ad-hoc networks that have varying node densities. These environments can be found in areas struck by disasters to urban city environments. Mobile nodes would not distribute evenly in such environments. Instead, the nodes would likely concentrate around specific landmarks or areas such as public squares with a high number of pedestrians while connecting roads and highways that have a low density of mobile nodes. The main contribution was the integration of a density based probabilistic scheme on AODV to reduce incurred routing overheads and improve packet delivery without significantly affecting the throughput and end to end delays. The proposed AODV–P (AODV–Probabilistic) protocol was compared against the AODV and OLSR protocols via simulation. The evaluation was based on three different environments: a high density, a variable density and a sparse density. The study showed that the AODV–P had a 28% higher data throughput than AODV and a 42% higher data throughput than OLSR at the three environments. Other performance parameters also showed that AODV–P performed better than AODV or OLSR in sub-optimal network environments due to the improved link lifetimes and lower routing overheads.