Research on the Learning Bidding Strategy of Bilateral Negotiation Transactions in Electricity Market

The bid of market players in bilateral negotiation transactions is often considered in combination with the opponent’s offer. In order to simulate the actual trading behavior, this paper uses Rubinstein bargaining theory to build a bidding model framework, which puts forward the market trading principles, refines the essential factors such as trading rounds, target price, the concession of the opponent, and bidding style of the primary market entity. This paper designs the discount factor of the model, combines with Bayesian theory to learn the offer of the opponent, dynamically updates and forecasts the bottom line of the opponent's bidding, and forms a learning model Bidding strategy. Through simulation calculation, this paper analyzes the mutual transactions of various bidding styles in the market, and reveals the relationship between the final transaction price and the combination of bidding styles. Learning a bidding strategy can promote the transaction price to evolve in a favorable direction, which is a stable form of market game. The results show the availability of the learning bidding strategy, which is helpful to transaction simulation and market analysis.