Thesis:

Title : A new hybrid clustering-based routing method for vehicular networks using the Dragonfly Algorithm and Tabu Search

Abstaract Vehicular Ad-hoc Networks (VANETs) face critical challenges in maintaining cluster stability and routing reliability due to high topological dynamics and node mobility. This research introduces a novel hybrid approach, named A new hybrid clustering-based routing method for vehicular networks using the Dragonfly Algorithm and Tabu Search (RVC-DATS), to optimize clustering and routing processes in these networks. In the proposed method, the Dragonfly Algorithm (DA) is employed for dynamic clustering and optimal Cluster-Head (CH) selection, maximizing intra-cluster cohesion by considering five behavioral criteria. To establish stable inter-cluster communication and avoid local optima during the routing process, the Tabu Search (TS) algorithm is utilized. Performance evaluation, conducted through MATLAB simulations in a highway scenario with 150 vehicles and a 250-meter communication range, demonstrates that RVC-DATS significantly outperforms baseline methods (PassCAR and GCAR). Specifically, it achieves a superior Packet Delivery Ratio (PDR) of approximately 66% and increases network throughput to nearly 900 KB/s. Furthermore, the reduction in standard deviation and standard error of node dispersion confirms the high stability of the proposed clustering structure under high-mobility conditions.

Keywords: Keywords: Clustering, Routing, Vehicular Networks, Dragonfly Algorithm, Tabu Search Algorithm