http://www.cnblogs.com/younghao/p/6696739.html
http://blog.sae.sina.com.cn/archives/5547
http://diyhpl.us/~nmz787/pdf/Human-level_control_through_deep_reinforcement_learning.pdf