Staying within the budget: Efficient probing strategies in opportunistic networks

In opportunistic networks, resource constrained mobile devices must continuously scan their environment for potential peers with whom they can exchange messages. This is an extremely energy consuming process and prior work has focused typically on how to adaptively changing the probing frequency by predicting future contact opportunities. However, without setting an energy budget for such probing activities beforehand, the standby time of smart phones cannot be guaranteed. In this paper, we provide a method to improve the probing efficiency given a certain energy budget. First, we discuss the characteristics of several real-life contact traces and motivate the need for an energy-efficient probing strategy in sparse scenarios. Subsequently, we propose an algorithm based on a Constrained Markov Decision Process, which maximizes the overall probing efficiency within a given energy budget. Simulating the algorithm on measured contact traces, we show that given a certain energy budget, the number of interesting contacts captured can be improved by 50% to 100% compared to a uniform probing policy.