A Multi-Agent System for Energy Management in a Dynamic and Open Environment: Architecture and Optimisation

Abstract

The increasing presence of smart and dynamic environments with IoT devices creates new challenges in energy optimisation such as handling environments’ dynamics and privacy concerns. In this paper, we aim at optimising energy consumption in such environments, exploiting building flexibility to reduce energy bills, while respecting user preferences as well as device constraints and addressing the complexity of the environment. We propose a multi-agent optimisation system based on Alternating Direction Method of Multipliers to solve the optimisation problem. Various evaluations of the proposal show significant energy cost savings, while addressing dynamics and preserving privacy.