Stochastic modeling of distributed,dynamic,randomized clustering protocols for wireless sensor networks

Distributed clustering architecture has been considered an effective and practical model to offer energyefficient, load-balancing, scalable, and robust communication for Wireless Sensor Networks (WSNs). In this paper, we compare and analyze various clustering schemes based on a comprehensive classification. We propose a bi-dimensional Markov chain model for analyzing a class of distributed, dynamic, and randomized (DDR) clustering schemes. With this model, we present extensive evaluation of stochastic properties of a representative DDR clustering scheme - Low Energy Adaptive Clustering Hierarchy (LEACH), in terms of the distribution of cluster number, the mean, the standard deviation and coefficient of variation of number of clusters. The results indicate that the number of clusters generated in LEACH-like DDR schemes is a random variable, which can not concentrate with in a narrow range of the optimal value. This variability in the number of clusters adversely affects the system lifetime.