In this seminar, Sorous h Karimi will give a talk about a Data-driven Trading Algorithm for S hort-Term Electricity Market.
Growth in the penetration of ren ewable energy sources makes electricity supply more uncertain and lea ds to an increase in the grid system imbalance and imbalance (real-ti me) price volatility. In some European countries, large-scale batteri es can be used to support transmission system operators (TSOs) in mai ntaining grid stability and earn profit, a practice called implicit b alancing. Model predictive control (MPC) strategies to exploit these implicit balancing strategies capture arbitrage opportunities but fai l to accurately model imbalance markets and face high computational c osts. Model-free reinforcement learning (RL) methods are fast to exec ute but require data-intensive training and usually rely on real-time and historical data for decision making. In this talk, I will presen t our proposed MPC-guided distributional RL method that combines the complementary strengths of both MPC and RL. The proposed method can e ffectively incorporate forecasts into the decision making process (as in MPC), while maintaining the fast inference capability of RL.
DESCRIPTION: Growth in the penetration of renewable energy sources mak es electricity supply more uncertain and leads to an increase in the grid system imbalance and imbalance (real-time) price volatility. In some European countries, large-scale batteries can be used to support transmission system operators (TSOs) in maintaining grid stability a nd earn profit, a practice called implicit balancing. Model predictiv e control (MPC) strategies to exploit these implicit balancing strate gies capture arbitrage opportunities but fail to accurately model imb alance markets and face high computational costs. Model-free reinforc ement learning (RL) methods are fast to execute but require data-inte nsive training and usually rely on real-time and historical data for decision making. In this talk, I will present our proposed MPC-guided distributional RL method that combines the complementary strengths o f both MPC and RL. The proposed method can effectively incorporate fo recasts into the decision making process (as in MPC), while maintaini ng the fast inference capability of RL. In this seminar, Soroush Kari mi will give a talk about a Data-driven Trading Algorithm for Short-T erm Electricity Market. END:VEVENT END:VCALENDAR