Selfish batteries vs. benevolent optimizers

The changing landscape of energy supply systems has spurred the development of new, distributed approaches for the energy management in distribution grids. Particularly, the dissemination of energy storage systems calls for new ways of coordination. Even though much has been written about agent-based systems in this context, a lack of coherence persists, posing challenges to the advancement of the field. This article offers a conceptual framework for analyzing and classifying multiagent-based energy management based on plural rationality theory, exploiting the close structural connection between multiagent and social systems. With the help of this framework, key contributions to the field were analyzed, accounting for diverse problem framings, objectives, and techniques. Furthermore, distinct strengths and weaknesses of the various strands found within the agent-based energy management literature were highlighted. One finding is that community-based approaches are underrepresented in the engineering literature despite being an active topic of investigation in neighboring fields of energy research. Giving more room to this perspective will not only have a likely impact on acceptance, but may also lead to new technical developments.