HENNEQUIN Sophie

Assoc. Prof. Dr. - Studies Director, Ecole Nationale d’Ingénieurs de Metz (ENIM) - Université de Lorraine 

September 2018 - present

- Taught industrial engineering in a French engineering school, and in an associated school in China,

- Developed and implemented decision making tools for companies,

- Managed 850 students (from bachelor to master level), 70 teaching colleagues, 30 research teaching colleagues and 120 teaching colleagues from industry.

My main research topics concern : the design/redesign and optimized operation of production systems and supply chains subject to various hazards and constraints, using both continuous approaches to facilitate analytical studies and discrete approaches to facilitate numerical studies. The applications developed to date concern the definition of production control strategies integrating product distribution and maintenance activities during the production system's operating phase. At supply chain level, reverse logistics and resource pooling activities (physical resources and logistics services) are also being studied, along with the consideration of suppliers. The new information and communication technologies developed as part of Industry 4.0, combined with these models, make it easier to dimension control strategies. The idea is to couple the two approaches (continuous and discrete) in order to propose simulation-based optimization methods that ensure optimal results. However, one of the main scientific locks remains the coupling between analytical and numerical models, which is very complex when systems evolve in real time (as trajectories in this case may be biased with respect to the initial trajectory studied with the continuous model). To ensure optimal operation, we need to be able to react and readjust in real time. However, we have developed methods to facilitate the link between numerical and continuous models.

The integration of the human factor in the models proposed (based on fuzzy models) enables better consideration of the interactions between the various players within a logistics or production system. To achieve this, agent-oriented approaches have been developed, where agents can be economic (game theory) and/or intelligent (multi-agent systems). From a methodological point of view, the objectives are to be able to justify the results obtained by proposing mathematical sensitivity analyses (based on methods such as infinitesimal perturbation analysis) while considering different evolution scenarios depending on the possible behaviors of the players, but also on possible modifications to the system and its environment. System dynamics are considered not only by the behavioral models defined, but also by the integration of organizational aspects and cooperation between different entities. Industry 4.0 technologies, such as IIoT, also make it possible to collect the information needed to feed the defined models and ensure that modifications are taken into account as quickly as possible. Combined with blockchains, IIoTs can then facilitate mutualization and symbiosis activities. The main scientific locks are linked to the limits of technological tools (e.g. blockchains ensure the security of exchanges, but are limited in capacity, the multiplicity of transactions is not possible, and consume a lot of energy for the block storage of data, IIoTs do not operate totally in real time, there is a time lag, and the multiplicity of technologies does not facilitate the definition of a global architecture). What's more, it's difficult to anticipate the various possible dynamics of evolution, as many events can occur at the same time, making it difficult to define reliable trajectories. The use of machine learning algorithms coupled to the models developed could provide a partial answer to these dynamic optimization problems, but since data collection can evolve considerably over time, reliable results are not guaranteed.

I have supervised 9 Ph.D. students, and 2 in progress.