3 questions for …
Mario Schmoltzi (35), Senior Manager Trading Strategies in the Trading & Optimization unit at STEAG, studied business administration with basic studies in business informatics und a focus on banking & finance in Hessen. He subsequently worked as an analyst for a US investment bank and then for an internationally operating British energy group. Since 2015 Mario Schmoltzi is in the employ of STEAG, where he brings his wide-ranging experience in electronic trading to bear in the Trading & Optimization unit.
What importance do computer-based algorithms have for STEAG’s T&O unit?
Without computer-based algorithms, electricity trading in its present form would not even be possible. For a long time, most of the trades have been handled by automated processes. The price fluctuations in the intraday electricity market, i.e., the very short-term trading in electricity deliveries for the same day, are tremendous. Trading is getting faster all the time, and more sophisticated. Decisions must be made at ever shorter notice – partly only a few minutes before the actual production and delivery of the electricity. Owing to this complexity, it is simply indispensable today to make use of computer-based algorithms in intraday trading.
How did you get the idea to integrate AI in the intraday trade?
Every day we move substantial sums scheduling and optimizing the dispatch of power generation systems. That itself is a highly dynamic process requiring a lot of experience. When the market environment and the plants and systems we market change, we frequently must analyze the new conditions in detail in order to adapt our activities accordingly. On the basis of deep reinforcement learning I have developed an algorithm that independently investigates its market environment by the trial and error method and deduces promising trading strategies from this investigation.
What is the potential of this algorithm for future trading?
AI algorithms will shape electricity trading in the future because they are able to adapt continuously to the market. This will have central importance especially in view of the changes in the energy market: The share of renewable energies will grow further, and the decentralizing of power generation will continue. If, for example, greater consideration is given to battery storage units in electricity trading, with this algorithm we can profitably market the flexibility they provide – and that in such a way that simultaneously the charge level of the battery is optimally taken into account.