PhD Position - Explainable and Traceable Hybrid Model Management for Digital Twins

Position Overview

We are seeking a motivated PhD candidate to contribute to the Engineering Digital Twins (EDT) program within Catalyst: the Reliable Hybrid Model Forge. The research focuses on developing explainable and traceable model management techniques for hybrid digital twins.

Digital twins are increasingly used across domains such as automotive, aerospace, manufacturing, medicine, robotics, and energy systems to monitor and optimize complex systems. They rely on multiple models to understand current system behavior and predict future performance. These models are often expressed using different modeling languages and paradigms, a situation commonly referred to as multi-paradigm modeling.

Recently, hybrid modeling approaches have become essential in digital twin development. These approaches combine physics-based models, which provide structural understanding and causal relationships, with machine learning models capable of learning patterns from operational data. While hybrid modeling improves predictive capabilities and adaptability, it also introduces new challenges related to model management, synchronization, traceability, and runtime reliability.

This PhD aims to develop methods that enable reliable, explainable, and traceable management of hybrid models throughout the lifecycle of digital twins.

Research Focus

Hybrid digital twins rely on the orchestration of heterogeneous models that evolve differently over time. Physics-based models may evolve due to design updates or changes in system constraints, while machine learning models adapt continuously based on newly available data.

Managing such heterogeneous and evolving models raises several key challenges:

• Heterogeneity of modeling paradigms and tools
  Digital twin models may be expressed in different languages and simulation platforms, requiring mechanisms to unify their semantics and support coordinated execution.

• Traceability and version management
  Updates to physics-based models and machine learning components must remain traceable in order to support debugging, validation, and safety-critical certification processes.

• Explainability of hybrid models
  Hybrid model behavior must remain interpretable to ensure trust in the predictions and decisions derived from the digital twin.

• Runtime stability and co-simulation reliability
  Digital twins often rely on co-simulation environments where models with different levels of fidelity may be dynamically switched to balance accuracy and computational latency.

This research will investigate model management (MoM) techniques based on model federation to address these challenges. By abstracting and unifying the semantics of heterogeneous models, model federation can support coordinated simulation, traceability of model evolution, and reliable integration of hybrid models.

Key Responsibilities

• Conduct research on model management techniques for hybrid digital twins
• Develop frameworks supporting explainable and traceable orchestration of heterogeneous models
• Design mechanisms for versioning, synchronization, and traceability of hybrid models
• Investigate runtime validation methods to detect model drift and maintain simulation stability
• Extend existing model federation infrastructures implementing the proposed model management approach
• Evaluate the approach through industrial case studies
• Publish research results in international conferences and journals
• Participate in collaborative research activities within the EDT program

Research Environment

The PhD will be conducted primarily at Télécom Paris (Institut Polytechnique de Paris) and will involve collaboration with IMT Atlantique and ABB Bangalore.

The research will take place in a multidisciplinary environment combining expertise in:

• Model-Based Systems Engineering (MBSE)
• Model-driven engineering
• Digital twin architectures
• Hybrid modeling and simulation
• Industrial applications of model management

Facilities and Resources

• Access to model-driven engineering platforms and model federation tools
• Collaboration with industrial partners and real-world digital twin use cases
• Participation in international research networks on digital twin technologies

Qualifications

Required

• Master’s degree in Computer Science, Software Engineering, or related field
• Strong background in model-driven engineering or system modeling
• Interest in digital twins, simulation, and cyber-physical systems
• Programming experience (Java, Python, or similar)
• Good written and oral communication skills in English

Preferred

• Knowledge of Model-Based Systems Engineering (MBSE)
• Experience with simulation platforms or co-simulation frameworks
• Familiarity with machine learning integration in engineering systems
• Background in model management or model federation techniques

Application Process

Please submit the following documents:

1. Cover Letter explaining your motivation and research interests
2. Curriculum Vitae
3. Academic Transcripts
4. Short Research Statement (1–2 pages)
5. Contact information for two references

Contact Information

Main supervisor Supervisor: Dominique Blouin Institution: Telecom Paris Email: dominique.blouin@telecom-paris.fr

Co Supervisors Supervisor: Sylvain Guérin Institution: IMT Atlantique Email: sylvain.guerin@imt-atlantique.fr

For scientific questions regarding the PhD topic, please contact the main supervisor directly.

Funding and Benefits

• Duration: 3 years
• Salary: Competitive PhD stipend according to French standards
• Benefits: Social security, health coverage, research travel support
• Travel: Support for conference attendance and research collaborations

Selected References

• Amrani, M., Blouin, D., Heinrich, R., Rensink, A., Vangheluwe, H., & Wortmann, A. (2021). Multi- paradigm modelling for cyber–physical systems: A descriptive framework. Software and Systems Modeling, 20(3), 611–639. https://doi.org/10.1007/s10270-021-00876-z
• Von Rueden, L., Mayer, S., Sifa, R., Bauckhage, C., & Garcke, J. (2020). Combining Machine Learning and Simulation to a Hybrid Modelling Approach: Current and Future Directions. In M. R. Berthold, A. Feelders, & G. Krempl (Eds.), Advances in Intelligent Data Analysis XVIII (Vol. 12080, pp. 548–560). Springer International Publishing. https://doi.org/10.1007/978-3-030-44584- 3_43
• Rudolph, M., Kurz, S., & Rakitsch, B. (2023). Hybrid Modeling Design Patterns (No. arXiv:2401.00033). arXiv. https://doi.org/10.48550/arXiv.2401.00033
• Thummerer, T., & Mikelsons, L. (2025). Learnable & Interpretable Model Combination in Dynamical Systems Modeling (No. arXiv:2406.08093). arXiv. https://doi.org/10.48550/arXiv.2406.08093
• Zendehboudi, S., Rezaei, N., & Lohi, A. (2018). Applications of hybrid models in chemical, petroleum, and energy systems: A systematic review. Applied Energy, 228, 2539–2566. https://doi.org/10.1016/j.apenergy.2018.06.051
• Slater, L., Arnal, L., Boucher, M.-A., Chang, A. Y.-Y., Moulds, S., Murphy, C., Nearing, G., Shalev, G., Shen, C., Speight, L., Villarini, G., Wilby, R. L., Wood, A., & Zappa, M. (2022). Hybrid forecasting: Using statistics and machine learning to integrate predictions from dynamical models. Hydrometeorology/Modelling approaches. https://doi.org/10.5194/hess-2022-334
• Syauqi, A., Pavian Eldi, G., Andika, R., & Lim, H. (2024). Reducing data requirement for accurate photovoltaic power prediction using hybrid machine learning-physical model on diverse dataset. Solar Energy, 279, 112814. https://doi.org/10.1016/j.solener.2024.112814
• Hallak, Y., Blouin, D., Pautet, L., Saab, L., Laborie, B., & Mittal, R. (2024). Model Management at Renault Virtual Simulation Team: State of Practice, Challenges and Research Directions. Proceedings of the ACM/IEEE 27th International Conference on Model Driven Engineering Languages and Systems, 1005–1014. https://doi.org/10.1145/3652620.3688223
• Amrani, M., Mittal, R., Goulão, M., Amaral, V., Guérin, S., Martínez, S., Blouin, D., Bhobe, A., & Hallak, Y. (2024). A Survey of Federative Approaches for Model Management in MBSE. Proceedings of the ACM/IEEE 27th International Conference on Model Driven Engineering Languages and Systems, 990–999. https://doi.org/10.1145/3652620.3688221
• Biglari, R. (n.d.). Foundations for Self-Adaptive Abstraction and Approximation in Digital Twins with Real-time Requirements. PhD Thesis, University of Antwerp, 2025. https://hdl.handle.net/10067/2148220151162165141

About the EDT Program

The Engineering Digital Twins (EDT) program is a major French research initiative bringing together leading research institutions to advance the science and engineering of digital twin technologies. The program aims to develop the scientific foundations and engineering methods required for next-generation digital twin systems.