Metaheuristic Algorithms for Cluster Head Selection in Heterogeneous Drone Networks
Bildiriler > Tam Metin Bildiri
Yazarlar: Ege Erberk USLU, Züleyha AKUSTA DAĞDEVİREN, Orhan DAĞDEVİREN
2025 9th International Artificial Intelligence and Data Processing Symposium (IDAP), , , Uluslararası, Tam Metin Bildiri, ss 1-8
DOI: 10.1109/idap68205.2025.11222347
Özet
The rapid proliferation of autonomous drone swarms brings the challenge of maintaining efficient and lowenergy communication topologies. Unlike traditional sensor networks, selecting cluster heads in these dynamic, three-dimensional, and heterogeneous environments requires advanced multi-objective optimization to balance coverage, reliability, and minimal overhead. In this study, we propose three metaheuristic algorithms for cluster head selection in heterogeneous drone networks: Genetic Algorithm (GA), Ant Colony Optimization (ACO), and Particle Swarm Optimization (PSO). In three-dimensional environments with varying communication ranges, the proposed optimization techniques aim to minimize the number of cluster heads while preserving independence constraints and full coverage. In the experimental evaluation, a total of 720 experiments were conducted across four different network scales (30-100 drones) using basic and aggressive parameter configurations. Constraint management techniques and multiobjective fitness evaluation balancing stability, energy efficiency, and coverage were integrated into all algorithms. The variation in the number of cluster heads remains below 0.3 units (Kruskal-Wallis$p>0.84$), indicating statistically equivalent solution quality across the algorithms. Under aggressive settings (Cohen's$d>6.0$), ACO completes 100 -drone scenarios in 14.7 seconds, whereas GA/PSO require 98.6-99.5 seconds. GA achieves better convergence for predictable execution scenarios, while ACO is the most suitable option for real-time applications that demand computational efficiency. These results provide an empirical basis for algorithm selection in resource-constrained drone deployments.