Journal of Applied Sciences1812-56541812-5662orgz10.3923/jas.2008.3924.3930SadehJ.RahimiyanM.122008821In this study, a new method for optimum coordination
of overcurrent relays is proposed. In order to solve the over current
relays coordination problem, a Q-learning based cooperative multi-agent
system is designed and applied. In the proposed method, each relay based
on the Q-learning algorithm can act as an autonomous adaptive intelligent
agent to find the desirable setting according to each power system configuration.
In addition, by injecting the cooperation concept in the reward function
of the Q-learning algorithm, all relays cooperate each other to minimize
the overall operating time of protection system, while, they satisfy the
coordination constraints. The efficiency of the suggested technique is
illustrated by applying the method for coordination of overcurrent relays
in a typical power system.]]>Urdaneta, A.J., L.G. Perez, J.F. Gomez, B. Feijoo and M. Gonzalez,2000Sadeh, J.,2005Keil, T. and J. Jäger,2008Zeineldin, H., E. El-Saadany and M. Salama,2004Razavi, F., H.A. Abyaneh, M. Al-Dabbagh, R. Mohammadi and H. Torkman,2008So, C.W. and K.K. Li,2000Mansour, M.M., S.F. Mekhamer and N.E.S. El-Kharbawe,2007Zeineldin, H.H., E.F. El-Saadany and M.M.A. Salama,2006Sutton, R.S. and A.G. Barto,1998Kaelbling, L.P., M.L. Littman and A.W. Moore, 1996Urdaneta, A.J., H. Resterbo, S. Marquez and J. Sanchez,1996