Algorithmic Energy Trading
PhD candidate: Arco de Kort
In recent years, there has been a shift in power generation sources to more renewable energy sources. This shift to renewables has caused more uncertainty in the volume of energy traded in the short-term energy markets, as they depend on weather conditions. The uncertainty creates more volatile energy prices, making it more challenging to make viable trading decisions.
This research focusses on developing self-learning algorithms that can deal with these new challenges. We first start with investigating the underlying dynamics of the short-term energy markets and proposing a method for forecasting the future prices. The forecasting method is then used for our trading policy. However, due to the volatile markets the accuracy of the forecasts may not be constant in time. Therefore to deal with this problem, we apply optimal learning in our algorithm so it can take this accuracy of the forecast into account when making a trading decision.
In this project, we want to contribute to both academic literature in intraday electricity trading and practical applications in energy trading and risk management. Better forecasting and trading algorithms can help market participants operate more efficiently while transitioning to more renewable energy sources.
External supervisor: Konrad Werkner