Data-driven Decision-making Strategies for Electricity Retailers: A Deep Reinforcement Learning Approach

With the continuous development of the electricity market, the electricity retailers, as the intermediaries between producers and consumers, have emerged in some of the liberalized electricity markets. Meanwhile, the electricity retailer faces many increasingly significant challenges from the complexities and uncertainties in both the supply and consumption sides. This paper applies a data-driven decision-making strategy via Advantage Actor-Critic (A2C) and Deep Q-Learning (DQN) for the electricity retailers. The retailers' profits and consumers' costs are both taken into account. This study verifies that the applied data-driven methods can handle the decision-making problem as well as promote the profitability of retailers in the electricity market. Furthermore, A2C is more appropriate than DQN in our simulation. The effectiveness of the applied data-driven methods is validated by using real-world data.

Automated summary

This study presents a data-driven decision-making strategy using A2C and DQN for electricity retailers, improving profitability and validating effectiveness through real-world data validation.