Researching and developing safe reinforcement learning algorithms for multi-energy management that stochastically solves the optimal planning and dispatch problem in day-ahead, intra-day and real-time energy markets, so that existing initial control strategies are improved or can be learned autonomously without prior knowledge and this with high stability guarantees.
|Effective start/end date||1/01/20 → 31/12/24|
- multi-energy management
Flemish discipline codes
- Automation and control systems