Agent-based approach to handle business complexity in US wholesale power trading

This study documents the practicality of an agent-based approach by examining how two groups of agents handle business complexity related to power trading. Three important findings are identified in this research. First, the proposed approach can estimate fluctuations of electricity prices as well as other well-known methods such as neural networks and genetic algorithms. Second, multiple learning capabilities incorporated in adaptive agents do not have an advantage over limited learning capabilities in predicting the market price of electricity. Finally, a theoretical extension of multiple learning capabilities may have potential for developing the agent-based approach for power trading.