AI for REAL-world NETwork operation

Información del proyecto

AI4REALNET

Identificador del acuerdo de subvención: 101119527

DOI

10.3030/101119527

Fecha de la firma de la CE 29 Junio 2023

Fecha de inicio 1 Octubre 2023

Fecha de finalización 31 Marzo 2027

Financiado con arreglo a

• Digital, Industry and Space

Coste total

€ 3 999 976,25

Aportación de la UE

€ 3 999 976,25

Coordinado por

INESC TEC - INSTITUTO DE ENGENHARIADE SISTEMAS E COMPUTADORES, TECNOLOGIA E CIENCIA
Portugal

CORDIS proporciona enlaces a los documentos públicos y las publicaciones de los proyectos de los programas marco HORIZONTE.

Los enlaces a los documentos y las publicaciones de los proyectos del Séptimo Programa Marco, así como los enlaces a algunos tipos de resultados específicos, como conjuntos de datos y «software», se obtienen dinámicamente de OpenAIRE .

Resultado final

Documents, reports (7)

Exploitation plan and strategy phase 2

This report will include an updated exploitation plan (compared to version V1) aimed to provide an updated and detailed description of the exploitation strategy of the consortium.

Communication and dissemination plan

This report will establish the basis for the development of a common communication and dissemination, plan in the project.

Exploitation plan and strategy phase 1

This report will establish a first basis for the development of individual and common exploitation plan in the project, including IPR considerations and open-source strategy.

AI4REALNET framework and use cases

This report will provide a qualitative description of human real-world decision-making and requirements for AI-based decision systems and understanding how context shapes individual decision-making. Moreover, it will present the use cases description and key performance indicators.

Communication and dissemination monitoring phase 1

This report will collect the main results produced within the communication and dissemination activities of the project.

Evaluation and test protocols

Report describing an evaluation protocol for assessing i) the qualitative and quantitative KPIs for the technical performance and scalability, ii) the robustness and safety of the of the AI solutions, and iii) the optimal balance between AI and human in the selected use cases.

Position paper on “AI for operation of critical energy and mobility network infrastructures”

Position paper describing the AI4REALNET view on AI for the operation of critical infrastructures, including scientific research challenges.

Other (3)

Digital environment – Version 1

Base environment version of the open-source digital environments software.

Digital environment – Version 2 (consolidated environment with interoperability connectors & KPIs)

Consolidated open-source digital environment (interoperability connectors, additional use cases, extended KPIs) software.

AI fundamental blocks – beta release

Open-source software implementation of transparent and trustworthy knowledge-assisted, hierarchical and distributed AI.

Data Management Plan (2)

Data management plan V2

Report that will update V1 and describe how research data is being handled during and after the project. It will describe which data is being (and will be) collected, processed, or generated, methodologies and standards to follow, whether and how this data will be shared and/or made open, and how it will be curated and preserved. Furthermore, it will give consideration to the ethics issues raised by the proposed work, in particular, it will provide an updated description and assessment of the ethical procedures.

Data management plan V1

Report that will define how research data will be handled during and after the project. This plan will define procedures on how to handle, store and share datasets according to the FAIR principles. It will describe which data will be collected, processed, or generated, methodologies and standards to follow, whether and how this data will be shared and/or made open, and how it will be curated and preserved. Furthermore, it will give consideration to the ethics issues raised by the proposed work, in particular, it will provide a description of the ethical procedures in order to allow a Ethics Check during the Project Review.

Publicaciones

Conference proceedings (19)

Multi-Objective Reinforcement Learning for Power Grid Topology Control

Autores: Thomas Lautenbacher, Ali Rajaei, Davide Barbieri, Jan Viebahn, Jochen L. Cremer
Publicado en: Power Tech 2025, 2025
Editor: https://www.arxiv.org/abs/2502.00040
DOI: 10.48550/ARXIV.2502.00040

Fault Detection for agents on power grid topology optimization: A Comprehensive analysis

Autores: Malte Lehna, Mohamed Hassouna, Dmitry Degtyar, Sven Tomforde, Christoph Scholz
Publicado en: European Conference of Machine Learning 2024, 2024
Editor: N/A

Sub-optimal Experts mitigate Ambiguity in Inverse Reinforcement Learning

Autores: Riccardo Poiani, Curti Gabriele, Alberto Maria Metelli, Marcello Restelli
Publicado en: Neural Information Processing Systems (NeurIPS) 2024, 2024
Editor: https://arxiv.org/abs/2401.03857
DOI: 10.48550/ARXIV.2401.03857

Imitation Learning for Intra-Day Power Grid Operation through Topology Actions

Autores: Matthijs de Jong, Jan Viebahn, Yuliya Shapovalova
Publicado en: European Conference of Machine Learning 2024, 2024
Editor: N/A
DOI: 10.48550/ARXIV.2407.19865

A Conceptual Framework for AI-based Decision Systems in Critical Infrastructures

Autores: Milad Leyli-abadi, Ricardo J Bessa, Jan Viebahn, Daniel Boos, Clark Borst, Alberto Castagna, Ricardo Chavarriaga, Mohamed Hassouna, Bruno Lemetayer, Giulia Leto, Antoine Marot, Maroua Meddeb, Manuel Meyer, Viola Schiaffonati, Manuel Schneider, Toni Waefle
Publicado en: IEEE SMC 2025, 2025
Editor: https://zenodo.org/records/16905470
DOI: 10.48550/ARXIV.2504.16133

On the Definition of Robustness and Resilience of AI Agents for Real-time Congestion Management

Autores: Timothy Tjhay, Ricardo J. Bessa, Jose Paulos
Publicado en: IEEE Power Tech 2025, 2024
Editor: https://zenodo.org/records/15237788
DOI: 10.48550/ARXIV.2504.13314

State and Action Factorization in Power Grids

Autores: Gianvito Losapio, Davide Beretta, Marco Mussi, Alberto Maria Metelli, Marcello Restelli
Publicado en: European Conference of Machine Learning 2024 , 2024
Editor: N/A
DOI: 10.48550/ARXIV.2409.04467

User experience evaluation of an AI-based decision-support tool for power grid congestion management

Autores: Jan Viebahn, Abdullah Ayedh, Jonas Lundberg, Magnus Bång, Jeroen Keijzers
Publicado en: AHFE International, Human Interaction and Emerging Technologies (IHIET 2025), Edición -5, 2025
Editor: AHFE International
DOI: 10.54941/AHFE1006694

Generation of Power Network Operating Scenarios for an AI-friendly Digital Environment

Autores: Jose Paulos, Pedro R. Silva, Ricardo J. Bessa, Antoine Marot, Jerome Dejaegher, Benjamin Donnot
Publicado en: IEEE PowerTech 2025 Conference, 2025
Editor: https://zenodo.org/records/15184237
DOI: 10.5281/ZENODO.15184237

Continuous Assessment Driven Requirements Elicitation For Trustworthy AI Systems

Autores: Wolfgang, Stefani; Heitz, Christoph; Chavarriaga, Ricardo
Publicado en: ECML PKDD AI-SCI 2025 workshop, 2025
Editor: https://zenodo.org/records/17099001
DOI: 10.5281/ZENODO.17099001

Study Design and Demystification of Physics Informed Neural Networks for Power Flow Simulation

Autores: Milad Leyli-abadi , Antoine Marot , Jérôme Picault
Publicado en: ECML PKDD ML4SPS 2025 workshop, 2025
Editor: https://arxiv.org/abs/2509.19233
DOI: 10.48550/ARXIV.2509.19233

Power Grid Control with Graph-Based Distributed Reinforcement Learning

Autores: Carlo Fabrizio, Gianvito Losapio, Marco Mussi, Alberto Maria Metelli, Marcello Restelli
Publicado en: ECML PKDD 2025 ML4SPS workshop, 2025
Editor: https://arxiv.org/abs/2509.02861
DOI: 10.48550/ARXIV.2509.02861

Last-Iterate Global Convergence of Policy Gradients for Constrained Reinforcement Learning

Autores: Alessandro Montenegro, Marco Mussi, Matteo Papini, Alberto Maria Metelli
Publicado en: Neural Information Processing Systems (NeurIPS) 2024, 2024
Editor: https://arxiv.org/abs/2407.10775
DOI: 10.48550/ARXIV.2407.10775

Learning Topology Actions for Power Grid Control: A Graph-Based Soft-Label Imitation Learning Approach

Autores: Mohamed Hassouna, Clara Holzhüter, Malte Lehna, Matthijs de Jong, Jan Viebahn, Bernhard Sick, Christoph Scholz
Publicado en: ECML PKDD 2025, 2025
Editor: https://arxiv.org/abs/2503.15190
DOI: 10.48550/ARXIV.2503.15190

How does Inverse RL Scale to Large State Spaces? A Provably Efficient Approach

Autores: Filippo Lazzati, Mirco Mutti, Alberto Maria Metelli
Publicado en: Neural Information Processing Systems (NeurIPS) 2024, 2024
Editor: https://arxiv.org/abs/2406.03812
DOI: 10.48550/ARXIV.2406.03812

Pioneering roadmap for ML-driven algorithmic advancements in electrical networks

Autores: Jochen L. Cremer, Adrian Kelly, Ricardo J. Bessa, Milos Subasic, Panagiotis N. Papadopoulos, Samuel Young, Amar Sagar, Antoine Marot
Publicado en: IEEE ISGT Europe 2024, 2024
Editor: https://arxiv.org/abs/2405.17184
DOI: 10.48550/ARXIV.2405.17184

Centrally Coordinated Multi-Agent Reinforcement Learning for Power Grid Topology Control

Autores: Barbera de Mol, Davide Barbieri, Jan Viebahn, Davide Grossi
Publicado en: ACM e-Energy 2025, 2025
Editor: https://arxiv.org/abs/2502.08681

Applying Job Design Criteria for Effective Human-AI Collaboration

Autores: Samira Hamouche, Nerissa Dettling, Toni Waefler
Publicado en: AHFE International, Human Interaction and Emerging Technologies (IHIET 2025), Edición -5, 2025
Editor: AHFE International
DOI: 10.54941/AHFE1006695

The Supportive AI Framework: From Recommending to Supporting

Autores: Toni Waefler, Samira Hamouche, Andrina Eisenegger
Publicado en: Lecture Notes in Computer Science, Augmented Cognition, 2025
Editor: Springer Nature Switzerland
DOI: 10.1007/978-3-031-93724-8_22

Peer reviewed articles (1)

Human-AI interaction in safety-critical network infrastructures

Autores: Marco Mussi, Alberto Maria Metelli, Marcello Restelli, Gianvito Losapio, Ricardo J. Bessa, Daniel Boos, Clark Borst, Giulia Leto, Alberto Castagna, Ricardo Chavarriaga, Duarte Dias, Adrian Egli, Andrina Eisenegger, Yassine El Manyari, Anton Fuxjäger, Joaquim Geraldes, Samira Hamouche, Mohamed Hassouna, Bruno Lemetayer, Milad Leyli-Abadi, Roman Liessner, Jonas Lundberg, Antoine Marot, Maroua Meddeb, Viola Schiaffonati, Manuel Schneider, Thilo Stadelmann, Julia Usher, Herke Van Hoof, Jan Viebahn, Toni Waefler, Giacomo Zanotti
Publicado en: iScience, Edición 28, 2025, ISSN 2589-0042
Editor: Elsevier BV
DOI: 10.1016/J.ISCI.2025.113400