Agent-based Simulation Experiment for Monthly Centralized Bidding Market in China

Abstract

China's electricity market reform is undergoing an upsurge. Different market models and market rules are gradually released. As an essential tool to research complex bidding behavior of market player under different market mechanisms, agent-based simulation is widely adopted. In this paper, two market models that implemented in Guangdong province, China, during the past two years including price spread and uniform clearing are concluded and modeled. And Roth-Erev reinforcement learning algorithm is used to model players bidding behavior. Numerical case focuses on the influence of models on the market profit of players. Results show that market model and its parameter setting have considerable influence on profit of players in the short term. But the influence becomes slight in the long term.