Leveraging LLM Agents for Automated Optimization Modeling for SASP Problems: A Graph-RAG based Approach

TL;DR

This paper introduces MAG-RAG, a novel automated optimization modeling approach using a multi-agent structure and graph-based retrieval-augmented generation, specifically designed for sensor array signal processing problems, outperforming existing benchmarks.

Contribution

It presents a new RAG-based method with a multi-agent architecture tailored for SASP, addressing the limitations of prompt-based techniques in domain-specific applications.

Findings

Outperforms several AOM benchmarks on classical SASP problems

Effectively matches user queries with domain knowledge using Graph-RAG

Enhances modeling accuracy in sensor array signal processing

Abstract

Automated optimization modeling (AOM) has evoked considerable interest with the rapid evolution of large language models (LLMs). Existing approaches predominantly rely on prompt engineering, utilizing meticulously designed expert response chains or structured guidance. However, prompt-based techniques have failed to perform well in the sensor array signal processing (SASP) area due the lack of specific domain knowledge. To address this issue, we propose an automated modeling approach based on retrieval-augmented generation (RAG) technique, which consists of two principal components: a multi-agent (MA) structure and a graph-based RAG (Graph-RAG) process. The MA structure is tailored for the architectural AOM process, with each agent being designed based on principles of human modeling procedure. The Graph-RAG process serves to match user query with specific SASP modeling knowledge, thereby enhancing the modeling result. Results on ten classical signal processing problems demonstrate that the proposed approach (termed as MAG-RAG) outperforms several AOM benchmarks.