Evolving buyer's bidding strategies using game-theoretic co-evolutionary algorithm

Abstract

This paper presents a co-evolutionary algorithm for evolving bidding strategies for buyers in a reconstructed pool-type electrical power market. A demand-driven algorithm which aims to closely follow the individual demands while maintaining low locational marginal price has been implemented and analyzed in detail based on simulation results under different market scenarios. The algorithm has been tested on a simulated power market with 7 buyers and 20 sellers in IEEE 14 bus network. The simulation results suggest that the proposed demand-driven co-evolutionary algorithm is an effective learning algorithm which helps the buyers optimize their bidding strategy. A novel hybrid algorithm which combines the demand-driven algorithm with a game-like decision making process has also been implemented to improve the performance of this algorithm.