28.10.2025
Sonja Wogrin and Luca Santosuosso presented their latest research from the ERC project NetZero-Opt at the INFORMS Annual Meeting in Atlanta.
The INFORMS Annual Meeting is one of the world’s leading conferences in operations research and management science, bringing together researchers from around the globe to exchange the latest advances in data-driven decision-making.
Representing the IEE, Sonja Wogrin and Luca Santosuosso were contributing two presentations from the ERC-funded project NetZero-Opt, which develops advanced optimization methods for planning net-zero energy systems.
Sonja Wogrin presented “Structured Learning for Time Series Aggregation in Net-Zero Power Systems: from Simplex Theory to Scalable Optimization”, showcasing how machine learning can be integrated into the simplex-based framework to further accelerate computations. The approach uses learning models to predict active constraints and solve parallel subproblems, significantly reducing computation times without compromising accuracy.
In his talk “Towards Exact Time Series Aggregation for Scalable Long-Term Power System Planning,” Luca Santosuosso introduced a novel, exact time series aggregation (TSA) method based on Dantzig’s simplex algorithm. The approach drastically reduces the temporal dimensionality of long-term energy system models while preserving exact optimality. This allows annual models with 8,760 hourly steps to be compressed into only a few representative periods—without any loss of accuracy—thus enabling more scalable system planning.