Teaching LLMs to Plan: Logical Chain-of-Thought Instruction Tuning for Symbolic Planning
This work addresses the limited ability of LLMs to perform structured symbolic planning, which is crucial for automated AI planning systems, representing a significant advancement in bridging general reasoning with logical precision.
The paper tackles the problem of enhancing large language models' (LLMs) symbolic planning capabilities, particularly in domains using formal representations like PDDL, by introducing a novel instruction tuning framework called PDDL-Instruct that teaches logical chain-of-thought reasoning, resulting in up to 94% planning accuracy on standard benchmarks, a 66% absolute improvement over baselines.
Large language models (LLMs) have demonstrated impressive capabilities across diverse tasks, yet their ability to perform structured symbolic planning remains limited, particularly in domains requiring formal representations like the Planning Domain Definition Language (PDDL). In this paper, we present a novel instruction tuning framework, PDDL-Instruct, designed to enhance LLMs' symbolic planning capabilities through logical chain-of-thought reasoning. Our approach focuses on teaching models to rigorously reason about action applicability, state transitions, and plan validity using explicit logical inference steps. By developing instruction prompts that guide models through the precise logical reasoning required to determine when actions can be applied in a given state, we enable LLMs to self-correct their planning processes through structured reflection. The framework systematically builds verification skills by decomposing the planning process into explicit reasoning chains about precondition satisfaction, effect application, and invariant preservation. Experimental results on multiple planning domains show that our chain-of-thought reasoning based instruction-tuned models are significantly better at planning, achieving planning accuracy of up to 94% on standard benchmarks, representing a 66% absolute improvement over baseline models. This work bridges the gap between the general reasoning capabilities of LLMs and the logical precision required for automated planning, offering a promising direction for developing better AI planning systems.