Q-Learning with Dynamic Rewards Table Applied to the SONET/SDH Ring Problem | IEEE Conference Publication | IEEE Xplore