deeprl_for_atari_games

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.