An Energy-efficient Clustering Algorithm Based on Affinity Propagation and Improved K-medoids for Mobile Wireless Sensor Networks

Abstract

With the development of mobile terminal technology, how to reduce the network energy consumption in mobile environment has become a research hotspot in mobile wireless sensor network (MWSNs). In this paper, an energy-saving clustering algorithm is presented based on improved K-medoids and affinity propagation (AP) in MWSNs. Firstly, the AP algorithm is used to find the number and location of the initial cluster heads. After that, a new weight function is established by considering the communication distance, residual energy and moving speed of the nodes. Next, improved K-medoids method is used to find the optimized cluster heads and form the network topology based on the new weight function. Finally, the network enters the communication stage using greedy algorithm to transmit data. The simulation results display that the presented algorithm can reduce the energy consumption of the entire network and prolong the network lifetime. Compared with LEACH, LEACH-M and APSA, the network lifetime of the presented APEEA algorithm is extended by 719 rounds, 684 rounds and 90 rounds, respectively. Meanwhile, the performance improvement of the energy consumption can achieve 16.553 J (66.2%), 13.973 J (55.9%) and 3.544 J (14.2%), respectively.