Distributed Reactive Power Control

This chapter discusses two types of the reactive power control methods. The first method is based on the well‐known artificial intelligence algorithm, Q‐learning algorithm. The second method is based on the distributed sub‐gradient algorithm. In reinforcement learning method, two agents exchange information with each other only when their own buses are electrically coupled. In order to reach the goal of minimizing the active power loss and satisfy operational constraints at the same time, a distributed Q‐learning algorithm is implemented. Q‐learning algorithm circumvents the dilemma to analyze complicated power system models. Multi‐agent system‐based distributed solution for optimal reactive power control, which is based on a distributed sub‐gradient algorithm, is appropriate for distributed computing.