ChatPlanner: A Large Language Model Framework for Personalized Public Transit Routing

In three linesChatPlanner is a framework using fine-tuned LLMs with RAG to extract user preferences from natural language and integrate them into public transit routing optimization. Evaluated on 8 personas and 5 contexts, the system combines fine-tuning (output structure) and RAG (query-specific context) to identify solutions overlooked by existing planners.

Summary generated by Claude — human-verified