Energy-Aware Communities Using Mobile Clients

摘要

This article distinguishes two aims: (1) to showcase an approach to identify clusters of users of mobile devices on the basis of shared common attributes, and (2) to investigate a social network-based awareness campaign which shows significant promise to reduce the individual's as well as the group's energy consumption. One possible way of progressively driving down the power consumption of individuals is to improve real-time and fine-grained awareness of their energy signature (or carbon footprint) relative to their peers. In this paper, we explore the possibility to support such an awareness program using gossiping in a network of mobile clients. We identify and address key challenges, namely how to propagate required information and, critically, how to filter out meaningless and / or counter-productive comparisons by detecting (through machine-learning techniques) boundaries between user communities. Results of a preliminary simulation study designed for evaluating performance in terms of our framework's ability to identify suitable comparison groups and to subsequently drive down the number of false positives / negatives are presented.