K-centers Min-Max clustering algorithm over heterogeneous wireless sensor networks

This paper proposes a clustering algorithm for heterogeneous wireless sensor networks, addressing energy dissipation as a key issue. Energy dissipation required by sensor nodes to transmit data depends on the distance between sensor nodes and cluster heads or a base station. Clustering is one of the best techniques for reducing energy consumption and extending sensor network lifetimes. Desirable features of the proposed clustering algorithm include: adaptation to changes in sensor distribution; energy efficiency; localized and distributed data aggregation and decision making; immunity to partial damage; and self-recovery. It employs a smallest disc covering algorithm to achieve a minimum of the maximum distance between a cluster head and sensor nodes compared to k-means clustering. Lawson's multiplicative rule is used for the smallest disc covering algorithm. Our simulation demonstrates that the proposed algorithm takes 50.8% fewer iterations to converge for cluster formation, with 33.9% and 23.2% shorter maximum and average intra-cluster distances versus k-means clustering. Performance is also improved.