Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/DCAF4FD3-F4F8-4B92-8A06-78CDCE3D2407 | ||
| Year | 2021 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Stavros Orfanoudakis, "Developing an autonomous agent for automated electricity trading", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2021 https://doi.org/10.26233/heallink.tuc.88411 | ||
| Appears in Collections |
The rise of renewable energy production, along with the recent popularization of electric vehicles, are gradually creating the need for a smarter electricity grid. Against this background, the PowerTAC trading agents competition and platform offers researchers (from academia and the industry alike) with effective means to test different business, market analysis, and market prediction strategies (potentially along with novel artificial intelligence algorithms), before even deploying them in the Smart Grid. In more detail, PowerTAC constitutes a multi-agent simulation platform for electricity markets, in which intelligent agents corresponding to electricity brokers compete with each other aiming to maximize their profits. Now, as AI researchers have found out the hard way time and time again, greediness almost never pays off in competitive multi-agent settings. In PowerTAC, too, agents that aim to take over a disproportionately high share of the market, might end up incurring financial losses due to being obliged to pay huge transmission capacity fees. Starting from this observation, we developed a novel trading strategy that aims to balance gains from controlling a sufficiently large part of the retail market, against the costs of paying high transmission capacity fees. We equipped TUC-TAC 2020, an agent that represented the Technical University of Crete in the PowerTAC-2020 international competition with this retail market strategy. Moreover, we developed a wholesale market strategy that utilized Monte Carlo Tree Search to determine TUC-TAC’s best course of action when participating in the market’s double auctions. Using these strategies, TUC-TAC was crowned the PowerTAC-2020 champion, competing against 7 other agents representing universities from 6 different countries. In this thesis, we present TUC-TAC’s 2020 strategy in detail; and also conduct an extensive post-tournament analysis, in order to draw important lessons regarding the strengths and weaknesses of the various strategies used in the PowerTAC-2020 competition.
Copyright © DIAS 2013
