GRaSP：地理分布式RAG的延迟感知优化
原标题：GRaSP: Latency-Aware Optimization for Geo-Distributed Retrieval-Augmented Generation
AI 摘要
该论文提出 GRaSP，一种面向地理分布式 RAG 的延迟感知重排序层，在语义相关性基础上按网络延迟衰减并可选加入跨区出口成本惩罚，同时满足全局与分区域选取上限，无需重训练或重建索引。作者证明基础问题可被 O(NlogN) 贪心算法最优求解，仅在加入硬性出口预算等约束时变为 NP-hard。在 CloudLab 四区域 WAN 与 Natural Questions 语料上，GRaSP 在均匀放置下几乎无损地降低抓取延迟，在主题化放置下给出优于仅相关性、最近邻和延迟预算基线的召回-延迟帕累托前沿，并在四个 BEIR 基准上验证了泛化性。
正文节选
GRaSP: Latency-Aware Optimization for Geo-Distributed Retrieval-Augmented GenerationThanks: Accepted at SIGIR-AP 2026, December 13–15, 2026, Singapore. The definitive version is available at https://doi.org/10.1145/3848000.3850202. Abstract. Retrieval-Augmented Generation (RAG) is increasingly served across geo-distributed cloud and edge infrastructures, where wide-area latency, cross-region egress cost, and data-residency rules compete with retrieval quality. Conventional pipelines rank by sema