We hereby invite top-class researchers of any nationality, interested in developing a collaborative application for an EU-funded Marie Sklodowska Curie Action Postdoctoral Fellowship (MSCA-PF-2022) project, to conduct research at VITO and its partners for a period of 1-2 years starting in 2023.

The competitive fellowship opportunities are 100% funded and include living and mobility allowances.

MSCA-PF 2022 information

The successful candidate will primarily work on the Marie Curie EU funded project, and she/he will be integrated in the VITO AMO (Algorithm Modelling and Optimization) Research Group, by taking part in regular meetings and discussion groups. The researcher will be introduced in the team’s regional and international networks.


Successful candidates will be supervised by Dr. Carlo Manna (

Dr. Carlo Manna is a researcher in applied artificial intelligence for energy systems. He had experience as senior researcher in academia (University College Cork 2011-2017) and as R&D scientist in industry (ZF-Automotive AG 2017-2019). Hi is co-authored of more than 30 journals/peer-review proceeding publications listed in Scopus and 7 German/US patents applications.

Target start date

The EU informs the results on the MSCA-PF applications in February 2023. Successful candidates are expected to be available to start within the following two months and no later than summer 2023.

Fellowship description

Context and research challenge:

Recent years have seen an increasing interest in Demand Response (DR) as a means to provide flexibility and increasing the penetration of renewable resource sin the power system. At the aggregator level, where there is a multitude and variety of devices and appliances across the aggregator’s portfolio(building HVAC, EVs, water tanks etc), the process of control and scheduling is rendered infeasible without automating a big part, or the whole process.

Real-life DR systems are increasingly an aggregation of a large number of heterogeneous systems, users and devices. Modelling such heterogeneous system is a big challenge for the following reasons:

  • the lack of complete information and data from the different components;
  • the unavailability or the high complexity of the mathematical models for the processes to be controlled/optimized;
  • the traditional approaches (mainly from the mathematical optimization and control theory) results in complex and intractable problems as the number of devices and therefore the variables involved increase [1];
  • there are many sources of uncertainty (e.g. dynamic prices, user behavior, renewable resources etc.) which introduce additional complexities to the ones described to the previous points.

Moreover, the increasing integration of renewable resources in the electric power grid has increased the need for producers and virtual producers (i.e. aggregators) to offer/trade or correct their offer close to real-time (e.g. intraday/continuous intraday market) in order to better deal with the high uncertainty due to the availability of renewables.

For these reasons, fast stochastic optimization and control methods are needed. However, the current approaches (MILP, MPC etc.) cannot guarantee a good ‘anytime’ solution for real-time control and scheduling due to the aforementioned computational issues.

Recently, there has been a large body of literature in deep learning that has sought to approximate or speed up optimization and control models. As described in reviews on topics such as combinatorial optimization [2], the underlying idea is to combine Machine Learning methods (to cope with complex stochastic models) and model-based approaches (integrating the knowledge about the process). Following this line of research, several approaches have recently been proposed in the literature as reported in [3] [4] [5] [6] [7].

The aim of this project is to investigate and implement innovative machine learning based methodologies for fast optimization and control approaches, particularly focused to support aggregators to trade/offer energy flexibility close to real-time.


We view our work as part of the aforementioned set of approaches, drawing from that important inspiration: that employing structural knowledge about the optimization model is paramount to achieving both feasibility and optimality.

Along considering the current state-of-art approaches, we will also exploit relevant findings from related ongoing and past VITO research projects (e.g. Induflex, Coordinet, Rolecs etc.) and VITO PhD project findings [8],as starting point for our algorithms development.

At same time, we will investigate and select some potential case-studies, suitable to customize the developed algorithms in order to build realistic, business oriented proof-of-concepts, using ongoing VITO projects with industrial partners (e.g. OpenLab etc) and/or (e.g. Pecan Street, REDD etc).

With this call, we invite researchers to submit their resumé (including track-record) and a one-page project description, that will be the basis for selecting candidates with whom we will collaborate for developing a competitive MSCA-PF proposal.


[1] Jordehi AR. Optimisation of demand response in electric power systems, a review. Renew Sustain Energy Rev 2019

[2]Yoshua Bengio, Andrea Lodi, and Antoine Prouvost. Machinelearning for combinatorial optimization: a methodologicaltour d’horizon. European Journal of Operational Research, 2020.

[3]W. Dong, Z. Xie, G. Kestor and D. Li, 'Smart-PGSim: Using Neural Network to Accelerate AC-OPF Power GridSimulation,'SC20: International Conference for High Performance Computing, Networking, Storage and Analysis, 2020

[4]Kyri Baker. Learning warm-start points for AC optimal power flow. In 2019 IEEE 29th International Workshop on MachineLearning for Signal Processing (MLSP), pp.1–6, 2019.

[5]Misra, S., Roald, L.A., & Ng, Y. (2021). Learning for Constrained Optimization: Identifying Optimal Active ConstraintSets.INFORMS Journal on Computing.

[6]Chen, D., Bai, Y., Zhao, W., Ament, S., Gregoire, J. & Gomes, C.. (2020). Deep Reasoning Networks for UnsupervisedPattern De-mixing with Constraint Reasoning. In Proceedings of the 37th International Conference on Machine Learning(ICML) 2020.

[7] Donti, P.L., Rolnick, D., & Kolter, J.Z. (2021). DC3: A learning method for optimization with hard constraints.ArXiv,abs/2104.12225.

[8]J. Arroyo, C. Manna, F. Spiessens, L. Helsen, Reinforced Model Predictive Control (RL-MPC) for Building EnergyManagement, Applied Energy,2022 (accepted for publications).


This is in progress. KU Leuven or UGhent, for the moment.

Deadline application to VITO

Interested candidates should submit their resume (incl. track record) and a one-page note describing the project for which a Marie Curie grant will be applied, as soon as possible and no later than Friday 17 June 2022 17h Brussels time.

Check the main eligibillity criteria: Marie Sklodowska-Curie | VITO.

For any inquiries please contact Dr. Carlo Manna at

Deadline MSCA-PF 2022

Wednesday 14 September 2022 17h Brussels time.

Meer info:…


We invite applicants to propose a more detailed and focused research approach within the scope of this MSCA-PF Fellowship as a part of their application. We are primarily looking for experienced researchers who wish to use this period as an opportunity to further develop their research and skills, and to develop longer-term research collaborations with VITO and other institutions conducting research in the field.

The candidates as in principle must be eligible for a Marie Curie Postdoctoral Fellowship – please refer to the conditions to be set-out in the Horizon Europe MSCA-PF-2022 Work Programme, including taking into account the new MSCA Green Charter principles.

The successful candidate should have the following qualifications/skills:

  • A PhD in a relevant field as (but not limited to): engineering, computer science, applied maths, physics etc.;
  • A strong background in one or more of the following areas: machine learning and data science, mathematical optimization, control theory, operations research, artificial intelligence;
  • Good background/interest in energy, and especially in energy system operation and planning, demand response, renewables integration, energy policy and modelling;
  • Experience with python/matlab or similar packages used for machine learning optimization/control and/or simulation and modelling will be highly beneficial;

The following assets will be advantageous:

  • An excellent track record in research, necessary for being able to develop a competitive Marie Curie Fellowship application;
  • Already published relevant research work in prestigious scientific journals;
  • An open and cooperation-oriented nature, but with strong abilities for independent research work;
  • highly proficient in spoken and written English.


Initially, we offer assistance in developing competitive Marie Curie Postdoctoral Fellowship proposals.

Then, to successful applicants to the Marie Curie programme, we offer;

  • An exciting opportunity at VITO, the independent Flemish research organisation driven by the major global challenges. Our goal? To accelerate the transition to a sustainable world;
  • Participation in a dynamic professional research & innovation community;
  • Flexible working conditions;
  • An inclusive and friendly work environment;
  • On-boarding assistance and other services.