Measuring the shape of data
Magnus Botnan
The human brain easily recognizes the circular shape formed by a few points scattered around a circle. But how do you construct algorithms which detect such non-linear shape in data?
That is the topic of persistent homology, a central part of the field of topological data analysis, which has been applied to many areas of science and engineering. While persistent homology enjoys important theoretical properties, it often fails to capture the shape of noisy data. The goal of this project is to understand the mathematical foundations of a generalization which addresses such shortcomings.
Power Network Optimimization in the Age of Climate Extremes
Alessandro Zocca
This project addresses the growing challenge of extreme weather events, such as floods, heatwaves, and windstorms, which seriously threaten the reliability of energy networks. These low-probability high-impact events are spatially correlated and can cause widespread physical and economic damage. Traditional methods fall short in assessing the resilience of these complex networks when facing these events. Combining applied probability, graph theory, and stochastic optimization, we develop new models and strategies for enhancing the resilience of these networks. We focus on finding cost-effective and flexible network reconfiguration strategies that operators can use to prevent and mitigate damage during such extreme weather events.