A Large Language Model-Based Agent for Automated Bidding Strategy Generation in Electricity Markets

As electricity markets become increasingly complex due to evolving market structures and the integration of renewable energy, optimizing bidding strategies for power generators has become essential to ensure competitiveness and efficiency. However, traditional bidding methods often struggle to adapt to market fluctuations and the inherent uncertainty of renewable energy generation. To address these challenges, we explore the potential of Large Language Models (LLMs) in developing intelligent and adaptive bidding strategies. This paper proposes a new agent for automated bidding strategy generation with systematic prompt engineering adopting the Chain of Thought (CoT) and auto-debug mechanism based on reflection. Experimental results demonstrate that our approach improves the accuracy, adaptability, and efficiency of automated bidding strategy generation in dynamic electricity markets. This paper highlights the potential of LLMs in energy trading, paving the way for more intelligent and autonomous decision-making agents.