filmov
tv
RL 5: Markov Decision Process - MDP | Reinforcement Learning

Показать описание
Markov Decision Process - MDP - Markov decision process process is a way to formalize sequential decision making process. Thus we can formalize reinforcement learning problem with finite markov decision process. There are 5 components of Markov decision process - the agent, the environment, the states, the actions and the rewards. The agents takes an action in the environment based on the current state of the environment. After every action the environment moves t[o another state. The agent receives a reward for it's action on the previous state. The goal of the agent is to maximize the total reward it receives in an episode or a specific number of steps.
Reinforcement learning tutorial series:
Reinforcement learning tutorial series:
Комментарии