A Wholesale Power Trading Simulator With Learning Capabilities

Abstract

Abstract--The US wholesale power market comprises a large commodity market. The growth in power trading is due to the ongoing deregulation policy of the electric power industry. Most of deregulation scenarios indicate a further separation of power production from transmission and retailing. The power production is opened to more competition. Unfortunately, the power trading mechanism is not clearly investigated in the level that we can predict a price change in the US wholesale power market. Such a price change in the US wholesale power market is explored from a simulation system with learning capabilities. Using the new intelligence system, we investigate the bidding strategies of traders in the wholesale power market and examine how the price change occurs under different economic and engineering environments.

Agent-Based Approach to Handle Business Complexity in U.S. Wholesale Power Trading

“…An agent-based application in the power industry is also found in [4]. The study incorporates several adaptive learning capabilities, based upon reinforcement learning (Erev and Roth [5, p.859]), in modeling various types of agents.…”

“…The study incorporates several adaptive learning capabilities, based upon reinforcement learning (Erev and Roth [5, p.859]), in modeling various types of agents. In the application, [4] has developed a new simulator in which agents can develop various bidding strategies to deal with the complexity of power trading.…”

“…From a theoretical perspective concerning power trading, the study [4] still has two major drawbacks. First, [4] does not document the practicality of the simulator.…”

“…First, [4] does not document the practicality of the simulator. Second, [4] does not describe how to incorporate "power law of practice" and "law of effect" in the simulator. The first law implies that adaptive agents need to include a learning curve that "tends to be steep initially and then flatter.…”

“…As an extension of [4], this study develops two groups of adaptive agents who investigate the dynamics of electric power trading. The first group consists of agents who are equipped with multiple learning capabilities.…”