Think Small, Plan Smart: Minimalist Symbolic Abstraction and Heuristic Subspace Search for LLM-Guided Task Planning

TL;DR

This paper introduces PLAHX, a minimalist symbolic abstraction and heuristic search framework that enhances LLM-guided task planning by reducing redundancy and search complexity, leading to higher success rates and lower token usage.

Contribution

The paper proposes a novel two-stage planning framework that combines minimalist symbolic abstraction with meta-heuristic search, improving efficiency and success in LLM-guided robotic task planning.

Findings

Improves success rate by 21.47% on average across domains.

Reduces token consumption by 13% compared to baselines.

Effective in complex domains like block stacking and robotic grasping.

Abstract

Reliable task planning is pivotal for achieving long-horizon autonomy in real-world robotic systems. Large language models (LLMs) offer a promising interface for translating complex and ambiguous natural language instructions into actionable plans. However, their probabilistic and opaque nature often leads to logically inconsistent or infeasible outputs. To address these limitations, recent frameworks combine LLMs with symbolic planners by first generating action models (Planning Domain Definition Language) and then applying heuristic search. Although promising, such systems still suffer from representation redundancy and exponential search complexity, often resulting in inefficient or overly long plans. To improve planning efficiency and effectiveness, we propose PLAHX (Planning from Language using Abstraction and Heuristic eXploration), a two-stage LLM-symbolic planning framework that integrates abstract symbolic representations with meta-heuristic subspace search in a parallel and iterative fashion. Rather than relying on verbose LLM-generated domain models, we introduce a minimalist symbolic abstraction pipeline that preserves semantic fidelity while eliminating redundancy. Our approach redefines LLM-symbolic planning not by making LLMs smarter, but by reducing the symbolic search space adaptively. Empirical results across four challenging domains, including block stacking and robotic mobile grasping, show that our approach improves the success rate by 21.47% on average, while reducing token consumption by 13% compared to state-of-the-art baselines.