Let's solve OpenAI Gym environments with some good old ways.

gym==0.18.0
numpy==1.21.5
imblearn==0.9.0
scikit-learn==0.24.0
xgboost==1.5.2

1. Play some game and gather state-action-reward data.
2. Distill a fine memory table based on reward.
3. (Optional) Use memory table, play and gather more data.
4. Get X-y data and train a model for control.
5. Use trained model, play and gather more data.

Will be added soon with details..