AMIES Award 2025 | Louis Bouvier: optimizing the return logistics of standardized packaging for automotive components

Sponsored by the learned societies Société Française de Statistique (SFDS), Société de Mathématiques Appliquées et Industrielles (SMAI), and Société Mathématique de France (SMF), the 2025 thesis prize was awarded during the 14^th edition of the Forum Entreprises & Mathématiques on Tuesday, October 7, 2025.

• Thesis title: Structured learning and combinatorial optimization: methodological contributions and inventory routing at Renault
• University of issuance: École Nationale des Ponts et Chaussées (IP Paris)
• Thesis supervisor: Axel Parmentier, researcher at CERMICS1  and professor at École Nationale des Ponts et Chaussées
• Company: Renault Group, specifically within the supply chain.
• Currently: Junior researcher and operational scientific manager, Co-Innovation Lab, École Nationale des Ponts et Chaussées

What motivated you to write a thesis related to the socio-economic world?

I was motivated by the idea of having a concrete impact, particularly by helping to reduce costs or CO2 emissions. I have always had a taste for theory applied to practice, an approach shared by my thesis supervisor. The CIFRE format allowed me to combine research and the business world, which suited me well because I didn't yet know exactly where I wanted to go in the medium term.

Can you tell us about your thesis topic?

My thesis focused on the return logistics of standardized packaging at Renault in Europe. Automotive components cannot be transported in standard cardboard boxes due to the risk of breakage. Renault therefore uses standardized reusable packaging that must be returned from the factories to the suppliers.

To reduce these costs, Renault has set up a system of shared packaging between its suppliers, which is managed centrally. The idea is that if a part is shipped across Europe, the packaging can be returned to the supplier closest to the factory rather than to its original supplier. This requires joint management of stocks in factories and at suppliers, as well as truck routes for transporting packaging, which corresponds to an inventory routing problem.

This problem is of a greater order of magnitude than those in the literature, both in terms of the number of sites and the number of types of packaging. In addition, as the journeys take several days, the arrival date depends on the order of customers on the routes, which is known as continuous-time inventory routing. Previous attempts by Renault's internal teams and service providers had been unsuccessful.

What were the main challenges you encountered during your research?

From a mathematical point of view, the first challenge was the size of the combinatorial optimization problem, on a European scale, which was an order of magnitude larger than anything else in the literature. The usual techniques did not work. I developed a mathematical heuristic combining mathematical programming formulations to create effective large neighborhoods, and integrated Bayesian machine learning methods to correct biases in the input data. It was also necessary to address the stochastic aspects of the problem, related to uncertainty about the release and demand for packaging.

From an application perspective, it was necessary to align the modeling and code with business needs, communicate the value of the tool to the various stakeholders, and support its implementation. These aspects proved essential to the success of the project.

How has your work benefited the socio-economic world today, or will it benefit it in the future?

The algorithm developed during my thesis has been put into production at Renault and has been in daily use since March 2023. This has resulted in savings of around €1 million per year and reductions in CO2 emissions of thousands of tons.

I have made the academic code, problem data, and solutions available as open source so that the scientific community can use them. The results have been published.

The methodological work in my thesis, at the intersection of operational research and machine learning, has led to new scientific collaborations. These methods, particularly those involving combinatorial optimization layers in neural networks, could potentially find applications in other sectors.

What advice would you give to young people who want to focus their mathematics research on practical applications?

I would advise them to choose applications that really interest and motivate them—the possibilities are endless! It's also important to find the right people to talk to, both in research and industry. Good supervision and committed industrial partners make things much easier.

I would also encourage them to dare to initiate collaborations and build relationships of trust with their industrial partners. These relationships are important for the project to have a real impact.

How do you see the role of mathematics in solving current societal issues?

Mathematics can contribute to societal issues, particularly environmental ones. In operational research, we develop tools to help make better decisions, whether it's to optimize supply chains, reduce CO2 emissions, or manage resources more efficiently.

More broadly, mathematics allows us to model phenomena and try to answer concrete questions in a rational way. It helps us quantify issues, explore different scenarios, and compare alternative solutions.

As a young researcher, how do you perceive the evolution of the link between academic research in mathematics and the socio-economic world?

On my level, I see an increase in the number of links between research and the socio-economic world. I believe these collaborations are beneficial because they can help fund research while generating a concrete impact for industry and society.

What are your plans for the future?

I want to continue combining applied research and mathematical consulting for industry and the public sector. I work at the Co-innovation Lab at the École Nationale des Ponts et Chaussées, where I have taken on the operational management of a research partnership with Renault. The aim is to continue developing mathematical solutions that have a concrete impact while promoting the school's research work to other companies and public actors.

• 1CNRS/ÉCOLE DES PONTS ET CHAUSSÉES