Double DQN: The Greedy Action
Beginner
The problem, from first principles
The greedy action is the one with the highest estimated value. Both networks in Double DQN use this rule, so getting its tie-breaking right matters for reproducible targets.
From theory to code
Implement greedy_action(q), returning the index of the largest action value.
Constraints
- Return a Python
int.
Hints
Hint 1
Use np.argmax, which breaks ties by the lowest index.
Keywords
research-papersreinforcement-learningdeep-q-learning
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.