A Genetic Algorithm-Based, Dynamic Clustering Method Towards Improved WSN Longevity

• Published in: Journal of network and systems management 2017, Vol.25 (1), p.21-46
• Main Authors: Yuan, Xiaohui, Elhoseny, Mohamed, El-Minir, Hamdy K., Riad, Alaa M.
• Format: Article
• Language: English
• Subjects: Algorithms; Clusters; Communications Engineering; Computer Communication Networks; Computer engineering; Computer Science; Computer Systems Organization and Communication Networks; Connectivity; Dynamics; Energy consumption; Genetic algorithms; Genetics; Information systems; Information Systems and Communication Service; Networks; Operations Research/Decision Theory; Remote sensors; Searching; Sensors; Wireless networks
• ISSN: 1064-7570; 1573-7705
• DOI: 10.1007/s10922-016-9379-7

The dynamic nature of wireless sensor networks (WSNs) and numerous possible cluster configurations make searching for an optimal network structure on-the-fly an open challenge. To address this problem, we propose a genetic algorithm-based, self-organizing network clustering (GASONeC) method that provides a framework to dynamically optimize wireless sensor node clusters. In GASONeC, the residual energy, the expected energy expenditure, the distance to the base station, and the number of nodes in the vicinity are employed in search for an optimal, dynamic network structure. Balancing these factors is the key of organizing nodes into appropriate clusters and designating a surrogate node as cluster head. Compared to the state-of-the-art methods, GASONeC greatly extends the network life and the improvement up to 43.44 %. The node density greatly affects the network longevity. Due to the increased distance between nodes, the network life is usually shortened. In addition, when the base station is placed far from the sensor field, it is preferred that more clusters are formed to conserve energy. The overall average time of GASONeC is 0.58 s with a standard deviation of 0.05.