mlsanityData to 14 Oct 2026

Language Models For Generalised PDDL Planning: Synthesising Sound and Programmatic Policies

arXiv 2508.18507 ↗First posted 25 Aug 2025cs.AIcited by 7

On the map: World models → World models for language agents

Abstract (arXiv)

We study the usage of language models (LMs) for planning over world models specified in the Planning Domain Definition Language (PDDL). We prompt LMs to generate Python programs that serve as generalised policies for solving PDDL problems from a given domain. Notably, our approach synthesises policies that are provably sound relative to the PDDL domain without reliance on external verifiers. We conduct experiments on competition benchmarks which show that our policies can solve more PDDL problems than PDDL planners and recent LM approaches within a fixed time and memory constraint. Our approach manifests in the LMPlan planner which can solve planning problems with several hundreds of relevant objects. Surprisingly, we observe that LMs used in our framework sometimes plan more effectively over PDDL problems written in meaningless symbols in place of natural language; e.g. rewriting (at dog kitchen) as (p2 o1 o3). This finding challenges hypotheses that LMs reason over word semantics and memorise solutions from its training corpus, and is worth further exploration.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Placed in World models for language agents, one of 11 directions in World models. Builds on 41 of its 84 references that are in the map. First posted on arXiv 25 Aug 2025.

Builds on

41 of 84 references.

References · 41

  1. 2017Proximal Policy Optimization Algorithmscited by 15,215
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  4. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  5. 2015Continuous control with deep reinforcement learningcited by 3,407
  6. 2013Playing Atari with Deep Reinforcement Learningcited by 3,274
  7. 2015Trust Region Policy Optimizationcited by 2,580
  8. 2016Asynchronous Methods for Deep Reinforcement Learningcited by 2,258
  9. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,071
  10. 2015Deep Reinforcement Learning with Double Q-learningcited by 1,745
  11. 2022CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesiscited by 885
  12. 2023LLM+P: Empowering Large Language Models with Optimal Planning Proficiencycited by 501
29 more
  1. 2022PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Changecited by 382
  2. 2023On the Planning Abilities of Large Language Models : A Critical Investigationcited by 312
  3. 2024Evolution of Heuristics: Towards Efficient Automatic Algorithm Design Using Large Language Modelcited by 268
  4. 2023Text2Motion: From Natural Language Instructions to Feasible Planscited by 224
  5. 2024Let's Think Dot by Dot: Hidden Computation in Transformer Language Modelscited by 191
  6. 2023Generalized Planning in PDDL Domains with Pretrained Large Language Modelscited by 139
  7. 2024LLMs Still Can't Plan; Can LRMs? A Preliminary Evaluation of OpenAI's o1 on PlanBenchcited by 94
  8. 2022Structured, flexible, and robust: benchmarking and improving large language models towards more human-like behavior in out-of-distribution reasoning taskscited by 56
  9. 2019Learning Domain-Independent Planning Heuristics with Hypergraph Networkscited by 50
  10. 2022Predicate Invention for Bilevel Planningcited by 49
  11. 2019ASNets: Deep Learning for Generalised Planningcited by 42
  12. 2025Beyond Semantics: The Unreasonable Effectiveness of Reasonless Intermediate Tokenscited by 41
  13. 2024VisualPredicator: Learning Abstract World Models with Neuro-Symbolic Predicates for Robot Planningcited by 39
  14. 2021Learning General Optimal Policies with Graph Neural Networks: Expressive Power, Transparency, and Limitscited by 38
  15. 2025LLMs as Planning Formalizers: A Survey for Leveraging Large Language Models to Construct Automated Planning Modelscited by 38
  16. 2018Learning Features and Abstract Actions for Computing Generalized Planscited by 35
  17. 2017Action Schema Networks: Generalised Policies with Deep Learningcited by 32
  18. 2024Thought of Search: Planning with Language Models Through The Lens of Efficiencycited by 32
  19. 2023Learning Domain-Independent Heuristics for Grounded and Lifted Planningcited by 31
  20. 2020Learning Generalized Relational Heuristic Networks for Model-Agnostic Planningcited by 22
  21. 2021Reinforcement Learning for Classical Planning: Viewing Heuristics as Dense Reward Generatorscited by 21
  22. 2024Return to Tradition: Learning Reliable Heuristics with Classical Machine Learningcited by 19
  23. 2025Classical Planning with LLM-Generated Heuristics: Challenging the State of the Art with Python Codecited by 18
  24. 2022Learning Sketches for Decomposing Planning Problems into Subproblems of Bounded Width: Extended Versioncited by 12
  25. 2021Discovering User-Interpretable Capabilities of Black-Box Planning Agentscited by 10
  26. 2022PG3: Policy-Guided Planning for Generalized Policy Generationcited by 8
  27. 2025Successor-Generator Planning with LLM-generated Heuristicscited by 7
  28. 2023Generalized Planning as Heuristic Search: A new planning search-space that leverages pointers over objectscited by 6
  29. 2025Leveraging Action Relational Structures for Integrated Learning and Planningcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 5.

7 papers.

  1. 2025Frontier Large Language Models Rival State-of-the-Art Plannerscited by 7
  2. 2025Learning Admissible Heuristics for A*: Theory and Practicecited by 2
  3. 2026On the Ability of Transformers to Verify Planscited by 2
  4. 2026LLM-Evolved Domain-Independent Heuristics for Symbolic AI Planningcited by 1
  5. 2026Property-Guided LLM Program Synthesis for Planningcited by 0
  6. 2026LLM-Evolved Pattern Generators for Optimal Classical Planningcited by 0
  7. 2026Learning Admissible Heuristics via Cost Partitioningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Leveraging Environment Interaction for Automated PDDL Translation and Planning with Large Language Models
  2. 2026Language Models as Higher-Order Planning Formalizers
  3. 2023Leveraging Pre-trained Large Language Models to Construct and Utilize World Models for Model-based Task Planning
  4. 2025An Extensive Evaluation of PDDL Capabilities in off-the-shelf LLMs
  5. 2025Generating Symbolic World Models via Test-time Scaling of Large Language Models
  6. 2026PDDLCoder: Agentic PDDL Generation for LLM-Assisted Symbolic Planning
  7. 2025End-to-end PDDL Planning with Hardcoded and Dynamic Agents
  8. 2024NL2Plan: Robust LLM-Driven Planning from Minimal Text Descriptions