Ethical Smart Grid: a Gym environment for learning ethical behaviours - Archive ouverte HAL Access content directly
Journal Articles Journal of Open Source Software Year : 2023

Ethical Smart Grid: a Gym environment for learning ethical behaviours

Abstract

The ethical-smart-grid package is a reinforcement learning (RL) simulator of a smart grid, based on Gym (Brockman et al., 2016). It focuses on learning ethical behaviours, i.e., behaviours that include ethical considerations, and are aligned with moral values. The simulated smart grid is composed of multiple agents representing inhabitants, which have to learn how to consume and distribute energy while taking into account several ethical considerations, such as the equity of comforts between inhabitants. This simulator is somewhat simplified compared to existing (co-)simulators (Karnouskos & De Holanda, 2009; Mihal et al., 2022; Oliveira et al., 2012; Vogt et al., 2018); it focuses instead on the ethical aspects. In other words, agents that have been trained on this environment could not be deployed in a real-world smart grid, yet this package still serves as a suitable environment for learning ethical behaviours.
Fichier principal
Vignette du fichier
10.21105.joss.05410.pdf (269.96 Ko) Télécharger le fichier
Origin Publisher files allowed on an open archive
Licence

Dates and versions

hal-04187666 , version 1 (25-08-2023)

Licence

Identifiers

Cite

Clément Scheirlinck, Rémy Chaput, Salima Hassas. Ethical Smart Grid: a Gym environment for learning ethical behaviours. Journal of Open Source Software, 2023, 8 (88), pp.5410. ⟨10.21105/joss.05410⟩. ⟨hal-04187666⟩
38 View
24 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More