Master's Thesis : Improvement of decision making for trading in wholesale electricity market
Moitroux, Olivier
Promotor(s) : Ernst, Damien
Date of defense : 25-Jun-2020/26-Jun-2020 • Permalink : http://hdl.handle.net/2268.2/9063
Details
Title : | Master's Thesis : Improvement of decision making for trading in wholesale electricity market |
Translated title : | [fr] Amélioration de la prise de décision pour les échanges sur le marché de gros de l'électricité. |
Author : | Moitroux, Olivier |
Date of defense : | 25-Jun-2020/26-Jun-2020 |
Advisor(s) : | Ernst, Damien |
Committee's member(s) : | Mathieu, Sébastien
Gemine, Quentin |
Language : | English |
Keywords : | [en] trading |
Discipline(s) : | Engineering, computing & technology > Computer science |
Commentary : | / |
Funders : | / |
Research unit : | / |
Name of the research project : | Improvement of decision making for trading in wholesale electricity market. |
Target public : | Student General public |
Institution(s) : | Université de Liège, Liège, Belgique |
Degree: | Master en ingénieur civil en informatique, à finalité spécialisée en "intelligent systems" |
Faculty: | Master thesis of the Faculté des Sciences appliquées |
Abstract
[en] It is common practice for risk-averse industrial companies to reduce their exposure to the volatile prices of the spot market by securing a base load supply on the year-ahead electricity market. While many research efforts have been put in designing strategies to interact with several markets and assets, the case of small industrial consumers bound to a block-size constrained click-by-click contract for the year-ahead market is overlooked in literature. This Master thesis seeks to explore this gap and aims at improving the purchase decision making process of such electricity consumers. Multivariate probabilistic forecasting is investigated as a mean to complement the trader's expertise. Compelling results are that year-ahead electricity prices expose random-like patterns which make future price inference extremely difficult. A comparison study of several time series model suggests that training global deep learning models on related time series noticeably improves the forecast accuracy but that simpler models produce better calibrated prediction intervals.
File(s)
Document(s)
Annexe(s)
Cite this master thesis
The University of Liège does not guarantee the scientific quality of these students' works or the accuracy of all the information they contain.