Binary Swarm Intelligence for Wireless Sensor Network Design

초록

Cluster-based hierarchical modeling is an effective approach for constructing wireless sensor networks (WSNs) that involves grouping nodes into clusters and electing cluster heads, with the collection of cluster heads in the network forming a connected dominating set. Obtaining an optimal dominating set is an NP-complete problem that has been tackled using various swarm intelligence (SI) algorithms, e.g., artificial bee colony (ABC), particle swarm optimization (PSO), and ant colony optimization (ACO) algorithms. This study compared the performance and optimal clustering design of WSNs using binary SI approaches dynamically designed to adapt to topological changes caused by any node in a WSN. Using SI, both the total communication distance and number of cluster heads can be simultaneously minimized to save energy in the sensor network. Simulation results indicate that binary ABC outperforms binary PSO, ACO, genetic algorithms, and simulated annealing. In addition, the best binary ABC results were found to be close to the optimal solutions obtained using CPLEX, with very small error percentages.

키워드