Appendix: ContinuousDoublyConstrainedBatch ReinforcementLearning