Iterative Formalization and Planning in Partially Observable Environments

TL;DR

PDDLego is a framework that enhances planning in partially observable environments by iteratively formalizing and refining environment representations using LLMs, without fine-tuning or trajectories.

Contribution

It introduces PDDLego, a novel iterative approach to formalize and improve planning in partially observable settings using LLMs without fine-tuning.

Findings

PDDLego improves planning success over end-to-end approaches.

It is robust against increasing problem complexity.

Domain knowledge from successful trials benefits future tasks.

Abstract

Using LLMs not to predict plans but to formalize an environment into the Planning Domain Definition Language (PDDL) has been shown to improve performance and control. While most existing methodology only applies to fully observable environments, we adapt to the more realistic and challenging partially observable environments without sufficient information to make a complete plan. We propose PDDLego, a framework to iteratively formalize, plan, grow, and refine PDDL representations by decomposing the environment and the goal into fully observable episodes. Without finetuning, in-context exemplars, or trajectories, PDDLego improves planning success and exhibits robustness against problem complexity compared to end-to-end approaches. We also show that the domain knowledge captured after a successful trial can benefit future tasks.