WSNs play an important role in monitoring and responding to dynamic events in applications such as disaster management, industrial automation, and environmental monitoring. Since event detection has the highest priority in a WSN, it is very crucial regarding timely response, energy efficiency, and system reliability. However, traditional threshold-based detection methods often lack flexibility and always consume large amounts of energy. SI algorithms, motivated by natural collective behaviors, offer robust and adaptive solutions for event detection in order to overcome the limitations of conventional methods. Performance analysis and comparison of three most powerful SI algorithms - DA, ACO, and PSO - for event detection in WSNs. Methods: The performance of these algorithms was analyzed for time efficiency, energy consumption, accuracy, and scalability using simulation-based experiments. The experiments simulate WSNs of different node densities and event characteristics on the widely used python and MATLAB platforms under identical conditions for each algorithm. Among them, the Dragonfly Algorithm outperformed ACO and PSO in time efficiency-66.48 seconds versus 85.10 and 77.03 seconds, respectively-energy consumption (0.000030 J vs. 0.000047 and 0.000035 J), and accuracy, 94.5% versus 92.1% and 93.2%, correspondingly. DA also has better scalability for different network sizes and performs consistently well for larger networks, up to 1024 nodes. ACO, though robust in path exploration, had the highest energy consumption and longest detection time, which seriously affects its scalability. PSO gave balanced performance, though it trailed DA significantly in key metrics. The Dragonfly Algorithm provides the most efficient and scalable solution for WSN event detection with high time efficiency, energy conservation, and accuracy. These results underline its applicability in real-world time-critical and energy-sensitive WSN applications.
Read full abstract