Relationship to Dynamic Programming

• agent posses perfect knowledge of the functions and
• Focused on how to compute the optimal policy with the least computational effort, assuming the environment can be simulated
• Q learning has NO knowledge of the functions and
• Focused on the number of real-world actions the agent must perform to converge to an acceptable policy
• in many practical domains , such as manufacturing problems, the costs in dollars and time of performing actions in the external world dominate computational costs