Large language models help simplify transportation planning tasks

Large Language Model Assisted Transportation Planning through a Differentiable User Equilibrium Layer

Artificial Intelligence

Summary

Transportation planning usually requires experts who understand complex traffic models, making it hard for regular decision makers to use. The authors created a system that combines a large language model and advanced traffic analysis methods to let non-experts ask natural language questions and get optimized transportation plans. This system includes a special layer that can differentiate traffic patterns, allowing it to improve plans based on goals like budget limits or observed traffic. The method was tested on two common tasks: adjusting travel times from traffic data and designing networks under budget constraints, showing potential to make transportation planning more accessible.

What this means in practice

  • For urban planners: Run natural language queries to calibrate traffic models and design road networks within budgets more easily.
  • For transportation modelers: Integrate differentiable user equilibrium layers with LLMs to accelerate optimizing traffic assignments and network configurations.
  • For logistics software developers: Build interactive transportation planning tools that simplify complex traffic optimization for non-expert users using AI-assisted frameworks.$Commercial implications: Enables development of user-friendly commercial transportation planning platforms that combine language models with optimization engines.

Authors

Eunseung Lee, Gyeongjun Kim, Keemin Sohn

Abstract

Transportation planning has traditionally required substantial expertise in network modeling, traffic assignment, and optimization. As a result, strategic analyses such as network design, capacity expansion, and model calibration remain inaccessible to non-expert decision makers. Recent advances in large language models (LLMs) provide an opportunity to bridge this gap by enabling natural-language interaction with complex transportation planning tools. However, existing LLM-based systems primarily serve as interfaces to simulation software and lack the ability to efficiently search, evaluate, and optimize transportation solutions. This study proposes an integrated human-AI transportation planning framework that combines a large language model, a model context protocol (MCP) interface, an open-source transportation planning engine (AequilibraE), and a differentiable user equilibrium (UE) assignment layer. The framework allows non-expert users to execute network analyses and improve solutions for optimization. A differentiable UE layer is developed by directly differentiating the Karush-Kuhn-Tucker (KKT) conditions of the Beckmann traffic assignment problem. The proposed framework is demonstrated through two representative planning tasks: calibration of travel time functions from observed traffic counts and network design under budget constraints. The examples illustrate how differentiable transportation models and LLM agents can jointly democratize transportation planning and substantially reduce barriers between domain expertise and transportation decision making.
View PDFOpen arXiv