EURO 2021 is the largest and most important conference for Operational Research and Management Science (OR/MS) in Europe organized by EURO – the European Association of Operational Research Society and the Hellenic Operational Research Society (HELORS). This year’s EURO conference took place in a hybrid format from 11-14 July 2021.
ATOM participated in EURO 2021 by presenting its energy disaggregation component. In particular, behavioural energy efficiency is among the top European priorities to reduce energy consumption. To access this potential, intelligent energy management and artificial intelligence are used to provide end-users with information and advices to increase efficiency. Such algorithms usually rely on vast amounts of data from IoT devices. However, most households are rarely in possession of equipment to such extent. For this purpose, ATOM introduces personalised apps to support behavioural energy efficiency. Energy disaggregation is used within the apps to provide appliance-level information, avoiding extensive use of smart meters. Still, to achieve adequate accuracy, algorithms either rely on preordained databases, which are not always applicable, or the collection of data for large periods, risking end-users’ losing interest in the process. In this study, we explored common energy disaggregation algorithms, like Factorial Hidden Markov Models, Combinatorial Optimization, Decision Tree Regression and Long Short-Term Memory, to inform ATOM’s behavioural decision support tools and establish an applied energy disaggregation framework that reduces required amount of data, total time of collection, and equipment cost to support intelligent energy management without sacrificing accuracy. A case in a typical Greek household was presented.
The program and book of abstracts of the conference can be found here:
