CES aims at building intuitive and interactive platforms for computational engineering problems that allow scientists and practitioners not only to understand and predict the behaviour of real systems but also to better capture the interaction between models and data and hence gain insights into unconventional and counter-intuitive phenomena.


CES targets the adaptive transition between hypothesis- and data-driven modelling, predictive simulation, control and quality assurance of complex (dynamical) systems governed by systems of differential-algebraic and/or partial differential equations applied among others to glacier evolution, energy harvesting, chemical and process engineering, medicine and surgery research, through the multi-scale design of lighter and stronger materials, and the modelling of organisms and diseases growth. As a second research strand, CES targets complex networks and their interaction with human behaviour, such as those arising in logistics, traffic, communication, energy, biological and social systems.

To achieve this goal, several challenges must be overcome, which are the core research directions of the Institute:

  • Data. Acquire, select, process and fuse data sets for phenomena and systems of interest.
  • Model. Select and evolve the proper mathematical models capturing the problem characteristics. Identify and quantify the most relevant parameters, given experimental evidence, in hypothesis-driven models. Shape data-driven and data-updated models using advanced statistics and artificial intelligence.  Master adaptation between hypothesis- and data-driven models. This includes adopting multi-scale and single-scale approaches, multi- or single-field problems and their interactions, and solving large-scale instances.
  • Simulate and Control. Discretise and control the computational complexity of the models and of the predictive simulations.  This will require working hand-in-hand with HPC developers, through co-design to optimise software and hardware for given computational needs, e.g. for (machine) learning algorithms and neural networks.
  • Assure Quality. Quantify, measure and control the effects of uncertainties and errors on quantities of interest to the modeller.
  • Visualise. Provide tools for interpreting and visualising phenomena in order to develop decision support systems for different application domains.

The Institute focuses on general methodological developments which are as application-independent as possible in order to streamline research and optimise open innovation and productivity.     

The research done at CES is primarily mathematical and data-driven modelling. We work in close synergy with (applied) mathematicians to ensure the mathematical rigour of the numerical methods we develop. We collaborate with computer scientists to create performing and robust computational techniques required for reliable analysis and control of complex systems. Finally, we include in the Institute an engineering flavour to guarantee that the theories and models developed are proven to provide societal and economic impacts. As such, CES is instrumental in building powerful and impactful interdisciplinary connections between engineering, computer science, mathematics, physics and other priority application areas in Luxembourg. Areas where we have already had impact include in particular: Space Science, Advanced Manufacturing and Materials, Robotics, Automotive, Renewable Energy, Process Engineering, Transport and Logistics as well as Neuro-degenerative Diseases.