Which strategy combines exploration and exploitation by choosing a random action with probability epsilon and the best known action otherwise?

Which strategy combines exploration and exploitation by choosing a random action with probability epsilon and the best known action otherwise?

• A. Exploration (Random) strategy
• B. Epsilon Greedy Strategy
• C. Monte Carlo method
• D. Temporal Difference (TD) method

Answer: (B)

Balancing exploration and exploitation is a key idea in reinforcement learning. The described strategy picks a random action with probability epsilon and the best-known action otherwise. This is the epsilon-greedy approach. It introduces exploration by occasionally trying random actions, which helps discover opportunities the current estimates might miss, while mostly exploiting by choosing the action with the highest estimated value to maximize rewards given what’s known. The epsilon parameter controls the trade-off: higher epsilon means more exploration, lower epsilon means more exploitation. Other options describe different concepts—pure exploration, sampling-based evaluation, or update rules—without the explicit policy of occasionally randomizing actions to discover better choices.

Balancing exploration and exploitation is a key idea in reinforcement learning. The described strategy picks a random action with probability epsilon and the best-known action otherwise. This is the epsilon-greedy approach. It introduces exploration by occasionally trying random actions, which helps discover opportunities the current estimates might miss, while mostly exploiting by choosing the action with the highest estimated value to maximize rewards given what’s known. The epsilon parameter controls the trade-off: higher epsilon means more exploration, lower epsilon means more exploitation. Other options describe different concepts—pure exploration, sampling-based evaluation, or update rules—without the explicit policy of occasionally randomizing actions to discover better choices.