Post-Doctoral Research Visit F/M Tractable Stochastic Optimization via Matrix Theory

European Commission; France; Sur place; EUR 35 000 - 45 000; Plein temps

Résumé du poste

A leading research institute in Lille is seeking a postdoctoral researcher to develop novel algorithms for stochastic optimization. The role involves collaboration with international experts and a focus on energy sector applications. The ideal candidate will have a Ph.D. and skills in programming, mathematical optimization, and scientific research. Competitive salary and additional benefits are offered.

Prestations

7 weeks of annual leave

Flexible organization of working hours

Public transport reimbursement

Professional equipment available

Social and cultural events

Qualifications

• Strong theoretical foundation and practical experience in mathematical optimization.
• Proficiency in stochastic optimization, particularly two-stage problems.
• Ability to conduct high-quality research with publications in reputable journals.

Responsabilités

• Conduct fundamental and applied research on the project’s core topics.
• Develop a robust software package for optimization frameworks.
• Design and execute comprehensive computational studies.

Connaissances

Mathematical Optimization

Numerical Linear Algebra

Programming Proficiency

Scientific Research

Formation

Ph.D. in Operations Research, Applied Mathematics, Computer Science, or related field

Outils

Python

NumPy

SciPy

Gurobi

CPLEX

COIN-OR

Description du poste

Inria, the French national research institute for the digital sciences

Organisation/Company Inria, the French national research institute for the digital sciences Research Field Computer science Mathematics Researcher Profile Recognised Researcher (R2) Country France Application Deadline 13 Sep 2025 - 00:00 (UTC) Type of Contract Temporary Job Status Full-time Hours Per Week 38.5 Offer Starting Date 1 Oct 2025 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Reference Number 2025-09264 Is the Job related to staff position within a Research Infrastructure? No

Offer Description

This position is situated within the framework of the TROMAT (TRactable stochastic Optimization through MAtrix Theory) project, a public-academic initiative with significant industrial collaboration.

The project is funded by the Gaspard Monge Program for Optimization (PGMO), a program launched by EDF (Électricité de France) and the Jacques Hadamard Mathematical Foundation (FMJH) to foster collaboration between academic and industrial researchers on major challenges in energy and complex systems. The research is specifically supported under the PGMO's IROE sub-program, which focuses on disciplinary approaches to optimization and operations research. This context ensures that the project's outcomes are both scientifically innovative and practically relevant.

The work will be carried out through an international collaboration between two leading research teams: the INOCS (Integrated Optimization with Complex Structure) team at Inria Lille - Nord Europe, France, and the Continuous Optimization Team at the RIKEN Center for Advanced Intelligence Project (AIP) in Tokyo, Japan. The postdoc will be hosted at the Inria center in Lille, joining the INOCS team, which specializes in developing new models and algorithms for optimization problems with complex structures, particularly in the energy sector and supply chain management. The project also involves a close partnership with researchers from EDF R&D, providing direct insight into the real-world challenges that motivate the research and a clear path for technology transfer and impact.

Assignments :

With the help of the project's principal investigators (Marius Roland, Pierre-Louis Poirion, and Wim van Ackooij), the recruited person will be central to advancing the core scientific objectives of the TROMAT project. The main assignment is to develop and analyze a novel matrix-theoretic perspective for scenario clustering in large-scale, two-stage stochastic optimization problems, with a particular focus on applications in energy networks.

The work will begin with a first phase, as outlined in the project's timeline. This includes conducting a comprehensive literature review to classify existing scenario clustering techniques within our proposed matrix framework, generating a diverse portfolio of test instances relevant to the energy sector (especially those of interest to EDF), and developing a robust, modular codebase for experimentation. Following this initial period, the postdoctoral researcher will take a leading role in driving the research forward along one of the project's primary axes: either developing a unified mathematical framework for clustering methods or exploring the use of numerical linear algebra techniques to construct computationally efficient model approximations.

The recruited person will be based at the Inria Lille - Nord Europe research center (Villeneuve-d'Ascq, France) and will be an active member of the INOCS research team. This project is a close international collaboration. The recruited person will be in connection with Dr. Pierre-Louis Poirion from the RIKEN Center for Advanced Intelligence Project in Tokyo, Japan, and Dr. Wim van Ackooij, a senior researcher at EDF R&D. This collaboration will involve regular virtual meetings to discuss progress and a one-month research visit to Tokyo to work directly with Dr. Poirion on the matrix-based and linear algebra aspects of the project. The collaboration with Dr. van Ackooij will ensure the research remains grounded in and relevant to practical challenges in the energy sector.

Responsibilities :

The person recruited is responsible for conducting the day-to-day research activities of the project. This includes the theoretical development, software implementation, design and execution of computational experiments, and rigorous analysis of the results. The postdoctoral researcher will be expected to work with a high degree of autonomy and will take initiatives for shaping the project's scientific direction in collaboration with the principal investigators. A key responsibility will be to lead the dissemination of research findings, which includes writing articles for publication in top-tier operations research journals and presenting the work at major international conferences. The researcher will be central to producing the project's main deliverables: a public database of instances, an open-source software package, and at least one high-impact journal paper.

Main activities:

• Conduct fundamental and applied research on the project's core topics. This involves an initial deep dive into the literature on scenario clustering and two-stage stochastic programming, followed by co-developing a specific research direction. You will design and implement novel algorithms that leverage a matrix-theoretic perspective to create tractable and high-quality approximations of complex optimization models.
• Develop a robust and modular software package for implementing and testing the proposed optimization frameworks. This includes creating modules for instance generation, constructing various clustering matrices, solving the resulting two-stage models, and establishing a comprehensive test bed for systematic experimentation and analysis.
• Design and execute a comprehensive computational study to validate the proposed methods. You will build and utilize a new database of two-stage stochastic programming instances, with a particular focus on real-world energy network problems relevant to our industrial partner, EDF. This involves analyzing the performance of the developed algorithms and comparing them against state-of-the-art benchmarks.
• Disseminate research findings through high-impact publications and presentations. You will be responsible for writing and submitting research articles to leading international journals in operations research and presenting your work at major scientific conferences. You will also prepare intermediate reports to share progress with project partners.
• Engage in international and cross-sectoral collaboration. You will work closely with all project members, including a planned one-month research visit to Tokyo to collaborate directly with Pierre-Louis Poirion at RIKEN. You will also interact with Wim van Ackooij to ensure the project's practical relevance to the energy sector.

Additional activities:

• Contribute to the creation of open and accessible research tools. This includes writing clear documentation for the developed software package and working with the team to make the code and the instance database available to the broader research community and our partners at EDF.
• Participate in the supervision of research interns or Master's students. As the project evolves, you will have the opportunity to mentor junior researchers who may contribute to specific applicative or exploratory aspects of the project, thereby developing your supervision and leadership skills.
• Assist in the development of future research directions and funding proposals. Building on the project's results, you will contribute to brainstorming and writing new research proposals for major funding agencies (such as the ANR), helping to shape the future of the research group's agenda.

The successful candidate must hold a Ph.D. in a relevant field such as Operations Research, Applied Mathematics, Computer Science, or a related discipline. A strong theoretical foundation and practical experience in the following areas are essential:

• Mathematical Optimization: A good understanding of stochastic optimization is appreciated, particularly with two-stage stochastic linear programs. Proficiency in linear programming (LP) and mixed-integer programming (MIP) is required.
• Numerical Linear Algebra: The project's core methodology relies on matrix theory. Therefore, the candidate should have a solid grasp of linear algebra concepts and matrix decomposition techniques (e.g., SVD, PCA, non-negative matrix factorization). Familiarity with randomized linear algebra methods is a significant plus.
• Programming Proficiency: Strong programming skills are crucial for the implementation and computational evaluation of the developed algorithms. The candidate must be proficient in a high-level programming language, with a strong preference for Python and its scientific ecosystem (NumPy, SciPy). Experience with optimization modeling languages and solvers (e.g., Gurobi, CPLEX, COIN-OR) is also expected.
• Scientific Research: A demonstrated ability to conduct high-quality research, evidenced by publications in reputable journals or conferences in the field of optimization or operations research.

Languages:

• English: A high level of proficiency in both written and spoken English is mandatory. English is the working language of the research team and the project's international collaboration. The candidate will be expected to write scientific papers and present their work at international conferences.
• French: Knowledge of French is appreciated as it would facilitate daily life and integration within the Inria Lille research center, but it is not a requirement for the position.

Languages FRENCH Level Basic

Languages ENGLISH Level Good

Additional Information

• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking and flexible organization of working hours
• Professional equipment available (videoconferencing, loan of computer equipment, etc.)
• Social, cultural and sports events and activities