A $1000 imes$ Faster LLM-enhanced Algorithm For Path Planning in Large-scale Grid Maps

2025-10-03

AI Summary

Traditional pathfinding algorithms (e.g., A*, Dijkstra) suffer from low search efficiency and high memory overhead on large-scale raster maps, while recent LLM-A* approaches exhibit poor planning efficiency and instability due to spatial illusions and low-quality waypoints. To address these limitations, this paper proposes iLLM-A*, an efficient path-planning algorithm integrating large language models (LLMs) with enhanced search mechanisms. Its core contributions are: (1) a lightweight A* optimization reducing computational redundancy; (2) an incremental learning–based waypoint generation framework ensuring high-quality, context-aware intermediate goals; and (3) an adaptive waypoint selection strategy dynamically balancing exploration and exploitation. Extensive experiments demonstrate that iLLM-A* achieves over 1000× average speedup versus LLM-A* (up to 2349.5×), reduces memory consumption by up to 58.6%, yields shorter paths, and significantly improves robustness across diverse map topologies.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Path planning in grid maps, arising from various applications, has garnered significant attention. Existing methods, such as A*, Dijkstra, and their variants, work well for small-scale maps but fail to address large-scale ones due to high search time and memory consumption. Recently, Large Language Models (LLMs) have shown remarkable performance in path planning but still suffer from spatial illusion and poor planning performance. Among all the works, LLM-A* cite{meng2024llm} leverages LLM to generate a series of waypoints and then uses A* to plan the paths between the neighboring waypoints. In this way, the complete path is constructed. However, LLM-A* still suffers from high computational time for large-scale maps. To fill this gap, we conducted a deep investigation into LLM-A* and found its bottleneck, resulting in limited performance. Accordingly, we design an innovative LLM-enhanced algorithm, abbr. as iLLM-A*. iLLM-A* includes 3 carefully designed mechanisms, including the optimization of A*, an incremental learning method for LLM to generate high-quality waypoints, and the selection of the appropriate waypoints for A* for path planning. Finally, a comprehensive evaluation on various grid maps shows that, compared with LLM-A*, iLLM-A* extbf{1) achieves more than $1000 imes$ speedup on average, and up to $2349.5 imes$ speedup in the extreme case, 2) saves up to $58.6%$ of the memory cost, 3) achieves both obviously shorter path length and lower path length standard deviation.}

Problem

• Accelerating path planning in large-scale grid maps
• Reducing memory consumption of LLM-enhanced path planning
• Improving path quality and computational efficiency

Innovation

• Optimizes A* algorithm for faster path planning
• Uses incremental learning to improve LLM waypoint generation
• Selects optimal waypoints for efficient path construction