Schinina, Roberto (2026) Convergence analysis of Exponential Weights algorithm in day-ahead electricity market auctions. Master's Thesis / Essay, Applied Mathematics.
|
Text
mAppM2026SchininaR.pdf Download (1MB) | Preview |
Abstract
In Day-Ahead energy auctions, bidders must bid strategically without prior knowledge of the market-clearing price or competitors’ decision-making process. While no-regret learning al- gorithms provide a framework to utilize post-auction data to adjust bidding choices, their long-term bid preferences remain unidentified. In this thesis, we address this gap by study- ing the asymptotic behavior of auction participants using the EW algorithm based on their cost functions, capacities, and strategy sets. For homogeneous players with two bid choices, we establish explicit conditions for the lowest-bid preference and show that if the aggregate capacities of the players equal the demand, then the highest bid is preferred. Moreover, we identify submartingale to supermartingale regime shifts when the aggregate capacity exceeds demand. Lastly, we extend this analysis to heterogeneous players, proving weight convergence across multiple strategic overlap scenarios.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Peypouquet, J.G. and Cherukuri, A.K. |
| Degree programme: | Applied Mathematics |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 02 Sep 2026 09:51 |
| Last Modified: | 02 Sep 2026 09:51 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/37441 |
Actions (login required)
 |
View Item |