Manuel Katholnigg

Brief description of my doctoral project

Who benefits from the energy transition increasingly depends on software. As tariffs grow more complex, algorithms controlling batteries, buildings, and factories determine who captures real savings and who does not. But to fully unlock these savings, forecasting and control must work as one: current algorithms treat them as separate steps, training prediction models to minimize statistical errors while ignoring the financial consequences downstream. My research closes this gap with algorithms that unify forecasting and control, optimizing directly for financial outcomes and transferring to new tariff structures without retraining. Running locally on the device, they also eliminate cybersecurity risks of centralized cloud control and convert the full potential of the energy transition into financial gains for households, businesses, and industry.