Learning Domain-Independent Heuristics for Grounded and Lifted Planning

2023-12-18 · Dillon Z. Chen, Sylvie Thiébaux, Felipe Trevizan

We present three novel graph representations of planning tasks suitable for learning domain-independent heuristics using Graph Neural Networks (GNNs) to guide search. In particular, to mitigate the issues caused by large grounded GNNs we present the first method for learning domain-independent heuristics with only the lifted representation of a planning task. We also provide a theoretical analysis of the expressiveness of our models, showing that some are more powerful than STRIPS-HGN, the only other existing model for learning domain-independent heuristics. Our experiments show that our heuristics generalise to much larger problems than those in the training set, vastly surpassing STRIPS-HGN heuristics.

📄 PDF Abstract BibTeX arXiv:2312.11143

Code (1)

DillonZChen/goose 공식 구현

Symmetry-Invariant Novelty Heuristics via Unsupervised Weisfeiler-Leman Features

2025-08-25 · Dillon Z. Chen arxiv

Novelty heuristics aid heuristic search by exploring states that exhibit novel atoms. However, novelty heuristics are not symmetry invariant and hence may sometimes lead to redundant exploration. In this preliminary repo…

Planning as Theorem Proving with Heuristics

2023-03-23 · Mikhail Soutchanski, Ryan Young

Planning as theorem proving in situation calculus was abandoned 50 years ago as an impossible project. But we have developed a Theorem Proving Lifted Heuristic (TPLH) planner that searches for a plan in a tree of situati…

Automated Theorem Proving

When both Grounding and not Grounding are Bad -- A Partially Grounded Encoding of Planning into SAT (Extended Version)

2026-03-19 · João Filipe, Gregor Behnke arxiv

Classical planning problems are typically defined using lifted first-order representations, which offer compactness and generality. While most planners ground these representations to simplify reasoning, this can cause a…

SYMBOLIZER: Symbolic Model-free Task Planning with VLMs

2026-04-20 · Sami Azirar, Zlatan Ajanovic, Hermann Blum arxiv

Traditional Task and Motion Planning (TAMP) systems depend on physics models for motion planning and discrete symbolic models for task planning. Although physics model are often available, symbolic models (consisting of …

Motion Planning

Reinforcement Learning for Classical Planning: Viewing Heuristics as Dense Reward Generators

2021-09-30 · Clement Gehring, Masataro Asai, Rohan Chitnis, Tom Silver 외

Recent advances in reinforcement learning (RL) have led to a growing interest in applying RL to classical planning domains or applying classical planning methods to some complex RL domains. However, the long-horizon goal…

reinforcement-learning Reinforcement Learning (RL)