Hajime Kimura Kazuteru Miyazaki Shigenobu Kobayashi Reinforcement Learning in POMDPs with Function Approximation. 152-160 1997 conf/icml/1997 ICML db/conf/icml/icml1997.html#KimuraMK97