Measuring the exercise of market power has always been a fundamental problem in electricity markets. According to a well-known economic definition, a seller’s market power is defined as “the ability to maintain prices profitably above competitive levels for a significant period of time”. This standard definition implies that market power has a fuzzy nature, which leads to the development of a fuzzy system for estimating the exercised market power. As understood from the definition, both duration and magnitude of market power play key roles in the measurement of the exercised market power. The fuzzy estimator gathers these two relevant aspects in a unified framework and uses them to assign a unique measure to the exercised market power. The fuzzy estimator is applied to measure each power supplier’s market power in an agent-based computational framework of the considered electricity market. In this agent-based simulation, the power suppliers use Q-learning to learn how to respond to competitors’ behavior and market conditions. The experimental results confirm the capability of the proposed fuzzy estimator to monitor and diagnose the exercise of market power.

Keywords