PhD Student in (multi-agent) reinforcement learning for health-aware control 1

IMOS

The Intelligent Maintenance and Operations Systems (IMOS) Lab at EPFL is looking for a motivated and out-of-the-box thinking PhD researcher, (100%, in Lausanne, fixed-term) starting in September or upon agreement.

Project description

The objective of this project is to develop novel methodologies based on (multi-agent) reinforcement learning for health-aware control of complex engineering systems. The research will focus on integrating system health and degradation dynamics into control strategies, enabling decision-making that jointly optimizes performance and long-term asset reliability.

The project will explore how reinforcement learning agents can learn control policies that account for operational objectives, physical constraints, and system aging, with a particular emphasis on multi-agent settings where multiple subsystems interact. Physics-informed models and data-driven approaches will be combined to ensure that learned policies are both efficient and consistent with underlying system dynamics.

Applications will include complex industrial and energy systems (e.g., wind turbines or other large-scale infrastructure), where control decisions have a direct impact on system lifetime.

This PhD position is part of an ERC Consolidator Grant, supporting cutting-edge research on health-aware control and intelligent maintenance of complex systems.

Profile

We are looking for a PhD candidate with a strong analytical background and an outstanding MSc degree in Mechanical Engineering, Computational Mechanics, Engineering Science, Physics, Applied Mathematics, or a closely related field.

You should have a solid foundation in machine learning (e.g., deep learning) and mathematical modeling, including experience with dynamical systems or differential equations. A strong interest in modeling physical systems and degradation processes (e.g., fatigue, damage accumulation) is expected.

Experience with graph neural networks or spatiotemporal models is highly desirable, as well as familiarity with physics-informed approaches that incorporate physical inductive bias into learning models.

 Knowledge of one or more of the following areas is considered a strong asset:

• Physics-informed machine learning and hybrid modeling approaches
• Computational mechanics, structural dynamics, or fatigue and damage modeling
• Signal processing and analysis of measurement data from physical systems
• Scientific machine learning or numerical methods for physical systems

Strong programming skills and the ability to work at the interface of machine learning and physics-based modeling of engineering systems are expected.

We expect the candidate to be self-driven, with strong problem-solving abilities and out-of-the-box thinking. Professional command of English (both written and spoken) is mandatory. 

Work Environment

EPFL is one of the most dynamic university campuses in Europe, ranks among the top 20 universities worldwide and offers an exceptional working environment with very competitive salaries.

The IMOS Lab (https://www.epfl.ch/labs/imos/ ) offers a highly motivating, interdisciplinary scientific environment with many opportunities to interact across projects and researchers, and maintains an excellent network of collaborations with industrial stakeholders and leading international universities.

Application process

Formal applications including:
·       a letter of motivation,
·       a CV of the candidate,
·       brief research statement (one page) describing your project idea in the field of reinforcement learning and health-aware control, making connections to your experience and related work from the literature,
·       transcripts of all obtained degrees (in English),
·       one publication (e.g. thesis or preferably a conference or journal publication, a link is sufficient) should be submitted via the application platform.

Further information on EPFL IMOS Lab can be found under: https://www.epfl.ch/labs/imos/

Shortlisted candidates will be invited to apply to one of the EPFL doctoral schools (e.g. EDRS, EDCE or EDEE). This parallel application process is necessary to be eligible for a PhD at EPFL.
 

Informations

Activity Rate : 100.00 

Contract Type: CDD

Duration: 1 year contract renewable (max. 6 years) 

Reference: 2199