Stage : Distributed Ledger Technology and Normative Multi-Agent for Smart Energy Management

Description de l'offre

  Advisors – Olivier Boissier , Philippe Calvez (ENGIE R&D)
·  Location – ENGIE CRIGEN 361, avenue du Président Wilson BP 33 9321 Saint-Denis La Plaine Cedex
·  Team – Connected Intelligence & Computer science and Intelligent Systems Dpt - R&D ENGIE LAB A2D2
Keywords: Multi-Agent Systems, Distributed Ledger Technology, Smart Energy Management

Summary
The global context of this research concerns the transformations of the global energy market due to the evolution of digital technologies. This global market is becoming a decentralized eco-system of several local and agile energy markets where prosumers trade the energy they produce as well as the one they consume.
In this context, Multi-Agent System (MAS) technologies are proposing promising directions: energy eco-system simulation, energy allocation optimisation, energy trading management on behalf of human users (individuals or energy stakeholders), virtual plant formation, etc.
Distributed Ledger System (DLS) is another promising direction of research in such a context. Issued from the extensions and generalizations of the blockchain technology (BT) based on Bitcoin, the current developments show that their applications are going beyond cryptocurrencies to address the management of smart contracts (e.g Scanergy Project).
While MAS technologies are targeted to the definition of autonomous agents and to their decentralized coordination thanks to coordination protocols, normative agent organizations, etc, the technologies supporting DLS are complementary. They are used to verify and store any transactions without relying on any central authority in control of the transactions. They share and make available to all nodes participating to the system, the information about every transaction ever completed. In such a context, smart contracts express agreements between two or more participants with contract terms corresponding to user-defined program executed in the DLS decentralized environment (e.g. ETHEREUM supporting a Smart Contract Language). Smart contracts can be used to reach agreements, to solve common problems. They can be enforced as part of transactions and are executed across the blockchain network by all connected nodes. for business process monitoring, smart contracts open new opportunities such as the support of Decentralized Autonomous Organizations (DAO).
In the context of a research convention ENGIE R&D/ Laboratoire Hubert Curien & Institut Henri Fayol at Mines Saint-Etienne, the objective of this research is to investigate how DLS Technology can be used to address the monitoring and enforcement of normative and coordination processes within agent organizations in open multi-agent system. The research questions to study are related to how these processes could be mapped onto a distributed ledger infrastructure. More precisely, we are interested in analyzing the interests of (i) integrating the automatic and immutable chain of transaction in the process monitoring facilities supporting the agent organizations, (ii) using smart contracts as control support of the collaborative and normative process in the context of open MAS, (iii) introducing mechanisms to ensure autonomy of the agents while also ensuring enforcement of their behaviour. These three questions have to be conducted considering the fact that at some point, several processes, organizations could be governing the agents. It is thus of first importance to address the management and coordination of multiple interdependent DLS at the same time. Interoperability and scalability are important features to consider.
The feasibility of the proposed models will be evaluated by prototyping a simplified smart energy management use case on top of the developed infrastructure.

Expected results

Theoretical
Expected theoretical results consist in state of the art analysis and model proposal for:
·  Integration of the automatic and immutable transaction history in the organization monitoring facilities of a normative multi-agent system
·  Using smart contracts as support of the collaborative and normative process control in the context of open agent organisations,
·  Introduction of mechanisms in the DLS infrastructure to ensure autonomy of the agents while also ensuring enforcement and trust.
These three analysis have to be conducted considering the management and coordination of multiple interdependent and interoperable chains/organisations at the same time.

Niveau d'étude :

Bac + 5 (Master 2 ou 3ème année d'école d'ingénieur)

Postuniversitaire (MBA, Doctorat)

Spécialité :

Informatique : Mathématiques - Computer Science / Mathematics

Informatique : cf description stage - See Internship description

Durée Proposée :

6 mois

Maître de stage :

Guillaume ARBOD

Localisation :

Site de la Direction Recherche et Technologies de ENGIE à Saint Denis - Stade de France (accès RER B, D, Métro ligne 13)