In reinforcement learning, exploration is the concept of evaluating all the possible options for an agent to maximize its rewards to reach its goal. Exploitation is the agent focusing on the best action known to the model in a state to get the reward. The primary challenge in reinforcement learning is knowing when to explore new states to discover alternate approaches or to exploit the states that the agent is already aware of in case those states lead to the goal.
Glossary of Artificial Intelligence (AI), Machine Learning (ML), and Big Data Terms
Exploration vs. Exploitation Tradeoff
