Learning in wireless sensor networks

Participants : Dana Marinca, Nesrine Ben Hassine, Pascale Minet, Selma Boumerdassi.

To guarantee an efficient and dynamic management of network resources and services we intend to use a powerful mathematical tool: prediction and learning from prediction. Prediction will be concerned with guessing the short-term, average-term and long-term evolution of network or network components state, based on knowledge about the past elements and/or other available information. Basically, the prediction problem could be formulated as follows: a forecaster observes the values of one or several metrics giving indications about the network state (generally speaking the network represents the environment). At each time $t$, before the environment reveals the new metric values, the forecaster predicts the new values based on previous observations. Contrary to classical methods where the environment evolution is characterized by stochastic process, we suppose that the environment evolution follows an unspecified mechanism, which could be deterministic, stochastic, or even adaptive to a given behavior. The prediction process should adapt to unpredictable network state changes due to its non-stationary nature. To properly address the adaptivity challenge, a special type of forecasters is used: the experts. These experts analyse the previous environment values, apply their own computation and make their own prediction. The experts predictions are given to the forecaster before the next environment values are revealed. The forecaster can then make its own prediction depending on the experts' "advice". The risk of a prediction may be defined as the value of a loss function measuring the discrepancy between the predicted value and the real environment value. The principal notion to optimize the behavior of the forecasters is the regret, seen as a difference between the forecaster's accumulated loss and that of each expert. To optimize the prediction process means to construct a forecasting strategy that guarantees a small loss with respect to defined experts. Adaptability of the forecaster is reflected in the manner in which it is able to follow the better expert according to the context.

In 2014, we applied on-line learning strategies to predict the quality of a wireless link in a WSN, based on the LQI metric and take advantage of wireless links with the best possible quality to improve the packet delivery rate. We model this problem as a forecaster prediction game based on the advice of several experts. The forecaster learns on-line how to adjust its prediction to better fit the environment metric values. A forecaster estimates the LQI value using the advice of experts. The model we propose learns on-line how to adapt to dynamic changes of the environment to compute efficient predictions. It presents a very good reactivity and adaptability. The simulations using traces collected in a real WSN based on the IEEE 802.15.4 standard have shown that the past time-windows which are effective for the prediction should have medium durations, about 200-400ms. The time windows durations less than 200ms do not give a good prediction, while durations larger than 400ms are efficient only in low variations environment. We note that these results strongly depend on the real traces, but the great advantage of the model is that it is self-adaptive to input traces profile. In this context, because of data normalization, the impact of loss functions is limited: entropy and square loss functions seem to give better and more stable predictions. Also, the experts prediction method should be adapted to traces profile. For low variation environment values, the average on past time windows is a good approximation. For high variation environment, a method predicting smoothed values close to minimum real values is more appropriate. Hence, the predicted values will be stabilized around the low values, avoiding estimations varying too much. Simulation results also show that for both types of experts (AMW and SES), the best expert depends on the phase considered. This is the reason why a forecaster is needed. Furthermore, the predictions of the EWA forecaster using SES experts are shown to be reactive and accurate. This combination minimizes the cumulated loss regarding the real LQI values, compared with any other combination such as EWA-AMW, BE-AMW and BE-SES, given by decreasing performance order.

Prediction and energy efficiency for datacenters

Participants : Dana Marinca, Nesrine Ben Hassine, Pascale Minet, Selma Boumerdassi.

The exponential development of Information and Communication Technologies (ICT) have led to an over consumption of services and data shared in networks. From computing in companies to unified communications through social networks and Internet of Things, the use of ICT a reach the highest level ever. The complexity involved by these different services reveals the limits of computing in companies and leads a majority of organisms to partially or completely host the management of there information system in data centers. The latter are larger and larger and are composed of buildings containing powerful computing equipments and air-conditionning systems. Data centers require a huge amount of energy. As an example, in 2014, the electric consumption of all date centers will be larger than 42 TWh, and after 2020 the CO2 production will be larger then 1.27 GTons, ie. more than the aeronautic industry (GeSI SMARTer 2020 report). These "frightening" figures led the research community to work on the management of energy consumption. Several tracks have been explored, among which the optimization of computation and load balancing of servers. At present, we work on tools dedicated to traffic prediction, thus allowing a better management of servers. Our work consists in modeling the traffic specific to data centers and apply different statistical prediction methods.