Exploring Reinforcement Learning Method in Bidding Strategy Development for Day-Ahead Electricity Market

Abstract

This paper introduces the detailed process of applying reinforcement learning to solve market participant bidding strategy problem. The process includes the setup of market clearing environment, reinforcement learning structure, and Q-learning algorithm. A comprehensive study on three specially designed problems demonstrates the Q-learning method can achieve significantly higher profit than the baseline method, which employs marginal cost as the offer price. The study provides insights to the learning process and the performance of Q-learning and demonstrates the performance varies with the changing condition of the environment, and tends to degrade with more complex patterns or random disturbances in the environment.

Meeting Name

12th IEEE PES Asia-Pacific Power and Energy Engineering Conference, APPEEC 2020

Department(s)

Electrical and Computer Engineering

Second Department

Computer Science

Research Center/Lab(s)

Center for High Performance Computing Research

Comments

Defense Advanced Research Projects Agency, Grant D18AP00054

Keywords and Phrases

Bidding Strategy; Electricity Market; Q-Learning; Reinforcement Learning

International Standard Book Number (ISBN)

978-172815748-1

International Standard Serial Number (ISSN)

2157-4839; 2157-4847

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2020 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.

Publication Date

13 Oct 2020

Share

 
COinS