Leveraging Knowledge Graphs and LLM Reasoning to Identify Operational Bottlenecks for Warehouse Planning Assistance

2025-07-23

AI Summary

To address the labor-intensive and inefficient manual bottleneck identification in high-dimensional, heterogeneous outputs of warehouse discrete-event simulation (DES), this paper proposes a knowledge graph (KG)–large language model (LLM) collaborative self-correcting reasoning framework. The method constructs a semantically rich, inferable operational KG and integrates iterative prompt engineering, automated Cypher query generation, and a self-reflection mechanism to enable multi-hop querying and dynamic subproblem decomposition—mimicking human diagnostic reasoning. Experimental results demonstrate near-perfect (≈100%) accuracy in identifying operational bottlenecks under equipment failure and process anomaly scenarios. Moreover, the framework significantly outperforms baseline methods on complex, relational diagnostic tasks, while reducing insight generation time by orders of magnitude. This work establishes a novel paradigm for automated root-cause analysis in intelligent manufacturing systems.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Data Mining & Knowledge Management: Intelligent Query Processing
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Analyzing large, complex output datasets from Discrete Event Simulations (DES) of warehouse operations to identify bottlenecks and inefficiencies is a critical yet challenging task, often demanding significant manual effort or specialized analytical tools. Our framework integrates Knowledge Graphs (KGs) and Large Language Model (LLM)-based agents to analyze complex Discrete Event Simulation (DES) output data from warehouse operations. It transforms raw DES data into a semantically rich KG, capturing relationships between simulation events and entities. An LLM-based agent uses iterative reasoning, generating interdependent sub-questions. For each sub-question, it creates Cypher queries for KG interaction, extracts information, and self-reflects to correct errors. This adaptive, iterative, and self-correcting process identifies operational issues mimicking human analysis. Our DES approach for warehouse bottleneck identification, tested with equipment breakdowns and process irregularities, outperforms baseline methods. For operational questions, it achieves near-perfect pass rates in pinpointing inefficiencies. For complex investigative questions, we demonstrate its superior diagnostic ability to uncover subtle, interconnected issues. This work bridges simulation modeling and AI (KG+LLM), offering a more intuitive method for actionable insights, reducing time-to-insight, and enabling automated warehouse inefficiency evaluation and diagnosis.

Problem

• Identify warehouse operational bottlenecks using Knowledge Graphs and LLM reasoning
• Transform raw simulation data into semantic Knowledge Graphs for analysis
• Automate inefficiency diagnosis in warehouses with adaptive AI-driven queries

Innovation

• Integrates Knowledge Graphs with LLM reasoning
• Transforms raw data into semantic Knowledge Graphs
• Uses iterative self-correcting LLM agent queries