1. Apprendre
    2. /
  2. Cours
    3. /
  3. Reinforcement Learning with Gymnasium in Python
Connected

Exercice

Defining epsilon-greedy function

In RL, the epsilon-greedy strategy is a balance between exploration and exploitation. This method chooses a random action with probability epsilon and the best-known action with probability 1-epsilon. Implementing the epsilon_greedy() function is crucial for algorithms like Q-learning and SARSA, facilitating the agent's learning process by ensuring both exploration of the environment and exploitation of known rewards, and this will be the goal of this exercise.

The numpy library has been imported as np.

Instructions

100 XP
  • Inside the function, write the suitable condition for an agent to explore the environment.
  • Choose a random action when exploring.
  • Choose the best action according to the q_table when exploiting.