Swarm Intelligence based optimization Of MANET cluster formation

The study of Mobile Ad-hoc Network remains attractive due to the desire to achieve better performance and scalability. This thesis describes a Swarm Intelligence inspired method of ad-hoc clustering to give a hierarchical structure to flat MANET. The proposed clustering algorithm derives it’s method of operation from ant behavior in their colonies. The algorithm originates from the findings of entomologists who, on observing the ant societies, have remarked that larvae and food are not scattered randomly about the nest, but in fact are sorted into homogeneous piles. The emergence of higher form and behavior from the collaborative operation of numerous entities of trivial intelligence makes this algorithm distributed, adaptive and scalable. The algorithm is devised to be independent of the MANET routing algorithm. Depending upon the context, the clustering algorithm may be implemented in the routing or in higher layers. In most applications of MANET, the node capabilities are constrained and the node function is heterogeneous during operation. This thesis identifies the various Roles played by nodes and uses that information to build a Role based routing model which takes the form of clusters. The dynamic formation of clusters helps reduce data packet overhead, node complexity, power consumption, and create multi-path routing. After cluster formation, specific nodes are elected as cluster-heads satisfying certain Roles and performance criteria. This thesis looks at the performance of the proposed clustering algorithm, when applied to random and pseudo-random mobility models. Studies on mobility models have indicated that temporally and spatially correlated mobility models with geographic restrictions are nearer to real life scenarios when compared to fully-random models. The proposed algorithm is proven to improve performance in pseudo-random mobility models.