Intelligent automatic generation control: Multi-agent Bayesian networks approach

A new intelligent agent based control scheme, using Bayesian networks (BNs), to design automatic generation control (AGC) system in a multi-area power system is addressed. Model independency and flexibility in specifying the control objectives, make the proposed approach as an attractive solution for AGC design in a real-world power system. The proposed control scheme is tested in simulation on a three areas power system and shows desirable performance. The results are also compared with the multi-agent reinforcement learning based AGC design technique.