Premium
A new rank‐order clustering algorithm for prolonging the lifetime of wireless sensor networks
Author(s) -
Mostafavi Seyedakbar,
Hakami Vesal
Publication year - 2020
Publication title -
international journal of communication systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.344
H-Index - 49
eISSN - 1099-1131
pISSN - 1074-5351
DOI - 10.1002/dac.4313
Subject(s) - wireless sensor network , cluster analysis , computer science , hierarchical clustering , data mining , algorithm , ranking (information retrieval) , key distribution in wireless sensor networks , sorting , sensor node , node (physics) , wireless , wireless network , computer network , artificial intelligence , telecommunications , structural engineering , engineering
Summary Energy efficient resource management is critical for prolonging the lifetime of wireless sensor networks (WSNs). Clustering of sensor nodes with the aim of distributing the traffic loads in the network is a proven approach for balanced energy consumption in WSN. The main body of literature in this topic can be classified as hierarchical and distance‐based clustering techniques in which multi‐hop, multi‐level forwarding, and distance‐based criteria are utilized for categorization of sensor nodes. In this study, we propose the approximate rank‐order wireless sensor networks (ARO‐WSNs) clustering algorithm as a combined hierarchical and distance‐based clustering approach. Different from absolute distance, ARO‐WSN algorithm utilizes a new rank‐order distance measure for agglomerative hierarchical clustering. Specifically, for each sensor node, we generate a ranking order list by sorting all other sensor nodes in the neighborhood by absolute distance. Then, the rank‐order distance of two sensor nodes is computed using their ranking orders. The designed algorithm iteratively group all sensor nodes into a small number of sub‐clusters. The results show that ARO‐WSN outperforms the competitive clustering algorithms in terms of efficiency and precision/recall. The lifetime of the network with the first node death criterion improved relative to LEACH, LEACH‐C, LEACH with fuzzy descriptors, and BPA‐CRP by 60%, 85%, 22%, and 18%, respectively, and with last node death criterion improved relative to K‐means, LEACH, LEACH‐C, and LEACH with fuzzy descriptors by 42%, 67%, 64%, and 24%, respectively.