Planning Anything with Rigor: General-Purpose Zero-Shot Planning with LLM-based Formalized Programming

TL;DR

This paper introduces LLMFP, a universal framework that uses large language models to formalize and solve complex planning problems as optimization tasks without task-specific training, achieving high success rates across diverse scenarios.

Contribution

The paper presents LLMFP, a novel, general-purpose approach leveraging LLMs for formalized, zero-shot planning as optimization, outperforming existing methods across multiple planning tasks.

Findings

Achieves 83.7% and 86.8% optimal rates on 9 planning tasks with GPT-4o and Claude 3.5 Sonnet.

Significantly outperforms baseline methods with 37.6% and 40.7% improvements.

Validated components through ablation experiments and analysis of success and failure cases.

Abstract

While large language models (LLMs) have recently demonstrated strong potential in solving planning problems, there is a trade-off between flexibility and complexity. LLMs, as zero-shot planners themselves, are still not capable of directly generating valid plans for complex planning problems such as multi-constraint or long-horizon tasks. On the other hand, many frameworks aiming to solve complex planning problems often rely on task-specific preparatory efforts, such as task-specific in-context examples and pre-defined critics/verifiers, which limits their cross-task generalization capability. In this paper, we tackle these challenges by observing that the core of many planning problems lies in optimization problems: searching for the optimal solution (best plan) with goals subject to constraints (preconditions and effects of decisions). With LLMs' commonsense, reasoning, and programming capabilities, this opens up the possibilities of a universal LLM-based approach to planning problems. Inspired by this observation, we propose LLMFP, a general-purpose framework that leverages LLMs to capture key information from planning problems and formally formulate and solve them as optimization problems from scratch, with no task-specific examples needed. We apply LLMFP to 9 planning problems, ranging from multi-constraint decision making to multi-step planning problems, and demonstrate that LLMFP achieves on average 83.7% and 86.8% optimal rate across 9 tasks for GPT-4o and Claude 3.5 Sonnet, significantly outperforming the best baseline (direct planning with OpenAI o1-preview) with 37.6% and 40.7% improvements. We also validate components of LLMFP with ablation experiments and analyzed the underlying success and failure reasons. Project page: https://sites.google.com/view/llmfp.