A list of my publications is accessible via Google Scholar, ORCiD, or the TU Delft. Some additional items, such as posters presented at conferences, are available from ResearchGate.

Below is an overview of the (primary) research areas I am currently working on. If you have great ideas for collaboration or a thesis on these topics, get in touch!

ALGORITHMS TO GENERATE ROBUST PLANNING ALTERNATIVES (MGA). Energy system models can support decisions on which technologies do deploy, and where, to achieve the energy transition. Ideally, they identify for us the 'optimal' solution to such decisions. Except, such an optimal solution is often not viable nor desirable in practice.

Luckily, many other technically-feasible and economically-comparable decision alternatives exist that are normally hidden from view. I work on optimisation methods that make these other options available, known as 'Modelling to Generate Alternatives', or MGA. By revealing these alternatives, we can find more viable, robust, consensus solutions. For example, see our application to the Italian energy transition.

APPROXIMATE MULTI-SCALE COMPUTING METHODS. Energy transition decisions often play out across scales, connecting neighbourhoods to national grids. Still, large-scale planning overlooks local technical constraints, producing misleading insights. Similarly, local system design disregards larger-scale system integration, missing out on synergies.

The critical integrating local and national planning requires extremely-high-resolution models, and innovation on their computational tractability. To make this possible, I research approximate computing methods, such as downscaling, machine-learning surrogate models or decomposition techniques. For instance, see the WindSPORES project.

HUMAN-COMPUTER INTERACTION METHODS. Novel, more powerful energy system planning methods, like the above MGA, require streamlining the more nuanced outputs they produce.

I work on human-computer interaction methods, such as interactive dashboards and human-in-the-loop algorithms, which facilitate exploring as well as re-computing the planning alternatives based on real-time preferences. See, for instance, this interview on our human-in-the-loop advancement of MGA methods.