Work

Cover Me: Mitigating Multi-Agent System Failure through Reinforcement Learning—A Technology Demonstration

Year: 2025

Type: conference-paper

Abstract: Agent systems are common subjects of investigation in power systems research.Multi-agent systems are robust and can provide guarantees with regards to convergence and solution quality, but cannot show resilience against circumstances not considered at design time.Deep reinforcement learning cannot per se give guarantees, but can adapt to unforeseen situations.The technology demonstration showcases a hybrid of a multi-agent system and a reinforcement-based agent, where the latter safeguards the former against failure.

DOI: 10.1145/3679240.3735501

Authors: Eric Veith, Arlena Wellßow, Torben Logemann, Emilie Frost

Institution: Carl von Ossietzky Universität Oldenburg

Language: English

Topic: Optimal Power Flow Distribution

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Affordable and clean energy, Industry, innovation and infrastructure

Open Access status: gold

OpenAlex ID: W4411345744

FWCI:

Cites: 5

Cited by:

Related to: 10