Understanding Retrieval-Augmented Generation (RAG): Enhancing Large Language Models