PIP-LLM: Integrating PDDL-Integer Programming with LLMs for Coordinating Multi-Robot Teams Using Natural Language

2025-10-26

AI Summary

To address the fragility of task decomposition, poor scalability, and low coordination efficiency in multi-robot collaborative planning, this paper proposes a natural language–driven two-tier planning framework. At the team level, robust task decomposition and allocation are achieved via dependency graph modeling; at the robot level, path planning and load balancing are jointly optimized using integer programming. The framework integrates large language models for instruction understanding, PDDL for symbolic task logic modeling, and mathematical optimization for numerical solution—enabling synergistic symbol reasoning and numerical optimization. Experiments demonstrate significant improvements in large-scale scenarios: plan success rate increases markedly, maximum and average travel costs decrease by 12.7%–28.4%, and load standard deviation reduces by 35.2%. The approach exhibits strong robustness, high scalability, and superior coordination performance.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Intelligent Robots: Motion and Path Planning
• Humans and AI: Planning and Decision Support for Human-Machine Teams

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Enabling robot teams to execute natural language commands requires translating high-level instructions into feasible, efficient multi-robot plans. While Large Language Models (LLMs) combined with Planning Domain Description Language (PDDL) offer promise for single-robot scenarios, existing approaches struggle with multi-robot coordination due to brittle task decomposition, poor scalability, and low coordination efficiency. We introduce PIP-LLM, a language-based coordination framework that consists of PDDL-based team-level planning and Integer Programming (IP) based robot-level planning. PIP-LLMs first decomposes the command by translating the command into a team-level PDDL problem and solves it to obtain a team-level plan, abstracting away robot assignment. Each team-level action represents a subtask to be finished by the team. Next, this plan is translated into a dependency graph representing the subtasks' dependency structure. Such a dependency graph is then used to guide the robot-level planning, in which each subtask node will be formulated as an IP-based task allocation problem, explicitly optimizing travel costs and workload while respecting robot capabilities and user-defined constraints. This separation of planning from assignment allows PIP-LLM to avoid the pitfalls of syntax-based decomposition and scale to larger teams. Experiments across diverse tasks show that PIP-LLM improves plan success rate, reduces maximum and average travel costs, and achieves better load balancing compared to state-of-the-art baselines.

Problem

• Translating natural language commands into feasible multi-robot plans
• Overcoming brittle task decomposition and poor scalability in coordination
• Optimizing robot-level travel costs and workload via dependency graphs

Innovation

• Integrates PDDL team planning with Integer Programming
• Translates commands into team plans then dependency graphs
• Optimizes robot task allocation using Integer Programming constraints