Building software agents for the power trading agent competition (PowerTAC)
Over the decades, average energy consumption has been on a gradual increase, as a result, there is a need for smart electrical grids to regulate electricity in a city to reduce to likelihood of sharp spikes in electricity demand, thus causing blackouts. From there, broker agents, the regulators of t...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/137947 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Over the decades, average energy consumption has been on a gradual increase, as a result, there is a need for smart electrical grids to regulate electricity in a city to reduce to likelihood of sharp spikes in electricity demand, thus causing blackouts. From there, broker agents, the regulators of the smart grids, are needed to employ competitive strategies in order to generate the most revenue when put into direct competition against other agents. The Power Trading Agent Competition (PowerTAC) provides a safe, competitive, and simulated environment for participants to test their broker agent strategies.
Hence, the focus of this paper is to describe the student’s development of a broker agent for the PowerTAC in detail, highlighting the agent’s Particle Swarm Optimisation implementation in the tariff market and presenting the performance of differing agent versions, as well as to describe on the agent’s Markov Decision Process implementation in the wholesale market. |
|---|