Performance Evaluation of Data Aggregation Functions using Markov Decision Processes

Abstract : Aggregation functions are intended to save energy and ca- pacity in Wireless Sensor Networks (WSNs), by avoiding unnecessary transmissions. Aggregation functions take benefit from spatial and/or temporal correlations to forecast or to compress the real data which are collected. Although several works have focused on data aggregation in WSNs, there is a lack of a formal unified framework that can compare several aggregation functions suitable for a given network topology, a given application and a target accuracy. We address this question in this paper by proposing a Markov Decision Process (MDP) that can help to evaluate the performances of aggregation functions. The performances are expressed using two new proposed metrics, which can assess the energy and capacity saving of aggregation functions. As illustrative examples, we use our MDP to evaluate and analyse the performances of basic aggregation functions (e.g. average) and more complex ones (time series, polynomial functions).
Complete list of metadatas

https://hal.inria.fr/hal-01251687
Contributor : Fabrice Valois <>
Submitted on : Wednesday, January 6, 2016 - 3:39:21 PM
Last modification on : Tuesday, November 19, 2019 - 12:13:23 PM

Identifiers

Citation

Jin Cui, Khaled Boussetta, Fabrice Valois. Performance Evaluation of Data Aggregation Functions using Markov Decision Processes. PE-WASUN '15 - the 12th ACM Symposium on Performance Evaluation of Wireless Ad Hoc, Sensor, & Ubiquitous Networks , ACM, Nov 2015, Cancun, Mexico. ⟨10.1145/2810379.2810384⟩. ⟨hal-01251687⟩

Share

Metrics

Record views

307