Ph. D. Project
Title:
Game theory-based algorithms to control intelligent manufacturing systems
Dates:
2020/02/20 - 2023/02/20
Description:
Nowadays, factories and supply chains are evolving to match the various challenges announced by the 4th industrial revolution (Industry 4.0, Factory of the
Future). This implies major changes, especially on the integration of a large number of intelligent and / or connected objects, as shown by the paradigms of
cyber-physical systems (CPS) Internet of Things (IoT). Thus, the various technological entities (products, resources) are equipped with new communication
and information technologies (ICT), allowing them to capture information, observe, decide and communicate. In addition to technology, one of the promises
of the 4th industrial revolution is related to the integration of humans, which belong to this set of intelligent objects, at all levels (strategic, tactical and
operational).
In terms of control, the last decade has shown that, despite the large number of hybrid control architecture designed to face industrial needs, no one have
been widely deployed in industry, mainly due of the presence of human factors. Recently, the works of Tsegay T. Mezgebe have focused on new control
algorithms, based on human-inspired relationships, such as negotiation (https://hal.archives-ouvertes.fr/CRAN/hal- 02384288v1) and consensus reaching
(https://hal.archives-ouvertes.fr/CRAN/hal-02126766v1). These works start from the postulate that human has to understand the system, which can only be
obtained if the system behaves like the human and requires a dynamic control function assignment.
The objective of this thesis is to explore the game theory to design control algorithms in the same context of Tsegay T. Mezgebe works on human-inspired
relationships between intelligent objects. The interest of game theory is that, on the one hand, there are similarities with multi-agent systems ( behavior) and,
on the other hand, several aspects allows to transpose human behaviors (role, rationality,... ), which adds the notion of human in the intelligent entity more
than the relationship aspect. Thus, the work may lead towards the premises of human virtual twin, designed in the multi-agent system to control an industrial
system.
The results of the Phd project will be demonstrated using Tracilogis, a technological platform based at ENSTIB/fiber and wood Campus
(http://www.enstib.univ-lorraine.fr/en/ research/presentation/). Thus, the domain of application will be in preference wood industry, textile and fibrous base
material.
Keywords:
CPS, MAS, GAME THEORY, Manufacturing Control Architecture
Conditions:
36 months
Campus France scholarship
Department(s): 
Eco-Technic systems engineering
Funds:
Campus France Grants