2026. Retrieval-Augmented Reinforcement Learning with Dynamic Deliberation Control for Knowledge-Intensive Large Language Model Applications. Journal of Advanced Artificial Intelligence Research. 5, 1 (May 2026).