An Adaptive Blind Algorithm for Energy Balanced Data Propagation in Wireless Sensors Networks ⋆ Pierre Leone 1 , Sotiris Nikoletseas 2 , and Jos´ e Rolim 1 1 Computer Science Department, University of Geneva, 1211 Geneva 4, Switzerland 2 Computer Technology Institute (CTI) and Patras University, P.O. Box 1122, 261 10, Patras, Greece Abstract. In this paper, we consider the problem of energy balanced data propagation in wireless sensor networks and we generalise previous works by allowing realistic energy assignment. A new modelisation of the process of energy consumption as a random walk along with a new analysis are proposed. Two new algorithms are presented and analysed. The first one is easy to implement and fast to execute. However, it needs a priori assumptions on the process generating data to be propagated. The second algorithm overcomes this need by inferring information from the observation of the process. Furthermore, this algorithm is based on stochastic estimation methods and is adaptive to environmental changes. This represents an important contribution for propagating energy bal- anced data in wireless sensor netwoks due to their highly dynamic nature. 1 Introduction Load balancing is a common important problem in many areas of distributed systems. A typical example is that of shared resources such as a set of processors, where it is of interest to assign tasks to resources without overusing any of them. A related but different aspect of load balancing appears in the context of sensor networks, where tiny smart sensors are usually battery powered: an important goal of data processing is to balance the total energy consumed among the entire set of sensors. However, limited local knowledge of the network, frequent changes in the topology of the network and the specifications of sensors, among others, make load balancing in sensors nets significantly different of classical load balancing in distributed systems. To our knowledge, these considerations were first pointed out in the field of sensor networks in [9]. In this paper the authors deal with the problem of devis- ing energy balanced sorting algorithms. In a subsequent paper [6] the authors ⋆ This work has been partially supported by the IST Programme of the European Union under contract numbers IST-2001-33135 (CRESCCO) and 001907 (DELIS). V. Prasanna et al. (Eds.): DCOSS 2005, LNCS 3560, pp. 35–48, 2005. c  Springer-Verlag Berlin Heidelberg 2005