CompassLLM: A Multi-Agent Approach toward Geo-Spatial Reasoning for Popular Path Query

2025-10-08

AI Summary

This paper addresses the problem of identifying popular paths in historical trajectory data. We propose a training-free, parameter-free, LLM-driven multi-agent framework. Methodologically, it employs a two-stage search–generation pipeline: (1) leveraging LLMs’ implicit spatial and graph-structural reasoning to locate high-frequency subpaths; and (2) orchestrating multi-agent collaboration to synthesize complete, plausible novel paths. Our key contribution is the first direct exploitation of LLMs’ latent spatial reasoning capabilities for trajectory pattern mining—enabling both dynamic adaptability and generalizable path generation without explicit supervision or fine-tuning. Experiments on real-world and synthetic datasets demonstrate that our approach significantly outperforms baselines in path retrieval accuracy, achieves competitive generation quality, and maintains controllable inference cost. The framework offers an efficient, scalable paradigm for urban planning, navigation optimization, and other spatiotemporal applications.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

The popular path query - identifying the most frequented routes between locations from historical trajectory data - has important applications in urban planning, navigation optimization, and travel recommendations. While traditional algorithms and machine learning approaches have achieved success in this domain, they typically require model training, parameter tuning, and retraining when accommodating data updates. As Large Language Models (LLMs) demonstrate increasing capabilities in spatial and graph-based reasoning, there is growing interest in exploring how these models can be applied to geo-spatial problems. We introduce CompassLLM, a novel multi-agent framework that intelligently leverages the reasoning capabilities of LLMs into the geo-spatial domain to solve the popular path query. CompassLLM employs its agents in a two-stage pipeline: the SEARCH stage that identifies popular paths, and a GENERATE stage that synthesizes novel paths in the absence of an existing one in the historical trajectory data. Experiments on real and synthetic datasets show that CompassLLM demonstrates superior accuracy in SEARCH and competitive performance in GENERATE while being cost-effective.

Problem

• Identifies most frequented routes from trajectory data
• Eliminates need for model training and parameter tuning
• Solves popular path queries using multi-agent LLM framework

Innovation

• Multi-agent framework leveraging LLMs for geo-spatial reasoning
• Two-stage pipeline combining SEARCH and GENERATE path operations
• Cost-effective solution requiring no model training or retraining