Agent-Orchestrated Adaptive RAG: A Comparative Study on Structured and Multi-Hop Retrieval
For researchers building RAG systems, this work provides empirical evidence that agentic enhancements are not universally beneficial, advocating for selective, cost-aware orchestration.
The paper introduces an Agent-Orchestrated Adaptive RAG framework with dynamic query decomposition and iterative retrieval, finding that query decomposition improves structured domain performance (overall score +0.04, MRR +0.17 on DevOps) but degrades multi-hop ranking, while reflection boosts citation accuracy at high latency cost.
Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by grounding their responses in external knowledge, but conventional pipelines rely on static, single-step retrieval that limits performance on complex queries. This paper presents an Agent-Orchestrated Adaptive RAG framework that introduces dynamic query decomposition, iterative retrieval, and a bounded self-reflective evaluation loop. We evaluate the system across two complementary datasets: a domain-specific DevOps knowledge base and the multi-hop reasoning benchmark MuSiQue. Using metrics that include overall score, citation accuracy, mean reciprocal rank, and topic coverage, we find that query decomposition yields consistent gains in the structured domain (overall score $+0.04$, MRR $+0.17$ on DevOps) but degrades ranking precision on the multi-hop benchmark, while the reflection mechanism improves citation accuracy at a substantial latency cost. These contrasting results show that agentic enhancements are not universally beneficial and must be applied selectively according to query and domain characteristics. Our findings argue for adaptive, cost-aware orchestration rather than uniformly aggressive reasoning pipelines.