A New Hybrid Algorithm for Determining the Optimal Number of Clusters Based on ICA, Hill Climbing and K-means Algorithms to Prolong WSN Lifetime

Imperialist Competitive Algorithm (ICA) is a Collective Intelligence algorithm whose performance is based on political, economic, and cultural interactions between countries. One type of the important problems solved by Collective Intelligence algorithms are clustering problems, which are applied to data classification, image segmentation, image desaturation, and clustering of nodes in Wireless Sensor Networks (WSNs). In this paper, a new hybrid algorithm is presented based on modified ICA, Hill Climbing, and K-means algorithms and its performance in WSNs is evaluated. Experiment results and comparing the performance of the proposed algorithm with that of a number of other algorithms show that the lifetime of the WSN whose nodes are clustered by the proposed algorithm is more than that of those whose nodes are clustered by the other algorithms.

Keywords:

Imperialist Competitive Algorithm , hybrid algorithm , Hill Climbing algorithm , K-means , clustering of nodes in WSNs