Toward PDDL Planning Copilot

2025-09-16

AI Summary

Large language models (LLMs) exhibit limited reliability in long-horizon planning within autonomous agent scenarios. Method: We propose the first PDDL-based planning coordination framework grounded in the Model Context Protocol (MCP), enabling LLMs to invoke external planning toolchains—via natural language instructions—to perform end-to-end tasks including syntax validation, planner selection, plan generation, solution verification, and execution simulation. Crucially, the framework requires no domain-specific fine-tuning and is compatible with any MCP-compliant LLM. Contribution/Results: Extensive experiments across three open-source LLMs demonstrate substantial improvements in planning success rate and accuracy over both tool-free baselines and GPT-4o. These results empirically validate that integrating dedicated symbolic planners significantly enhances LLMs’ long-horizon reasoning capabilities—highlighting both efficacy and broad applicability of tool-augmented planning for autonomous agents.

Technology Category

• Multiagent Systems: Multiagent Planning
• Planning, Routing, and Scheduling: Planning with Language Models
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large Language Models (LLMs) are increasingly being used as autonomous agents capable of performing complicated tasks. However, they lack the ability to perform reliable long-horizon planning on their own. This paper bridges this gap by introducing the Planning Copilot, a chatbot that integrates multiple planning tools and allows users to invoke them through instructions in natural language. The Planning Copilot leverages the Model Context Protocol (MCP), a recently developed standard for connecting LLMs with external tools and systems. This approach allows using any LLM that supports MCP without domain-specific fine-tuning. Our Planning Copilot supports common planning tasks such as checking the syntax of planning problems, selecting an appropriate planner, calling it, validating the plan it generates, and simulating their execution. We empirically evaluate the ability of our Planning Copilot to perform these tasks using three open-source LLMs. The results show that the Planning Copilot highly outperforms using the same LLMs without the planning tools. We also conducted a limited qualitative comparison of our tool against Chat GPT-5, a very recent commercial LLM. Our results shows that our Planning Copilot significantly outperforms GPT-5 despite relying on a much smaller LLM. This suggests dedicated planning tools may be an effective way to enable LLMs to perform planning tasks.

Problem

• Bridging LLMs' inability for reliable long-horizon planning autonomously
• Integrating multiple planning tools via natural language instructions
• Enabling syntax checking, planner selection, and plan validation without fine-tuning

Innovation

• Integrates planning tools via chatbot interface
• Uses Model Context Protocol for tool connectivity
• Supports planning tasks without domain fine-tuning