We implemented 6 different Q-network models, including:
- Linear Q-network without experience replay and target fixing.
- Linear Q-network with experience replay and target fixing.
- Linear double Q-network.
- Deep Q-network.
- Double deep Q-network.
- Dueling deep Q-network.