Research on Q-learning Algorithm of Multi-agent in Micro-grid Control System
Lu Bi · IEEE Software · 2013
This paper describes the structure of micro-grid control system firstly as well as the workflow of multi-agents, and then proposed a collaborative learning algorithm which is improved based on Q-learning algorithm used in the structure of multi-agent in micro- grid control system, to make it suitable for agent learning in the mixed environments. Finally, we make a simulation experiment in a micro- grid system in which IEEE 9 bus system is as a basic framework and the results show that the algorithm can restore the power to the steady state rapidly when the power fluctuates.