{"id":36530911,"url":"https://github.com/tasl-lab/LaMMA-P","last_synced_at":"2026-01-18T21:00:44.057Z","repository":{"id":293069111,"uuid":"949527263","full_name":"tasl-lab/LaMMA-P","owner":"tasl-lab","description":null,"archived":false,"fork":false,"pushed_at":"2025-09-23T04:44:13.000Z","size":775,"stargazers_count":21,"open_issues_count":6,"forks_count":5,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-09-23T06:24:27.726Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/tasl-lab.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-03-16T16:55:06.000Z","updated_at":"2025-09-23T04:44:17.000Z","dependencies_parsed_at":"2025-05-13T14:57:40.852Z","dependency_job_id":"d27a3a16-c357-4c58-b648-dc803cd06263","html_url":"https://github.com/tasl-lab/LaMMA-P","commit_stats":null,"previous_names":["tasl-lab/lamma-p"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/tasl-lab/LaMMA-P","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tasl-lab%2FLaMMA-P","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tasl-lab%2FLaMMA-P/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tasl-lab%2FLaMMA-P/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tasl-lab%2FLaMMA-P/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/tasl-lab","download_url":"https://codeload.github.com/tasl-lab/LaMMA-P/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/tasl-lab%2FLaMMA-P/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28550464,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-18T20:59:07.572Z","status":"ssl_error","status_checked_at":"2026-01-18T20:59:02.799Z","response_time":98,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-01-12T03:00:47.522Z","updated_at":"2026-01-18T21:00:44.042Z","avatar_url":"https://github.com/tasl-lab.png","language":"Python","funding_links":[],"categories":["A01_文本生成_文本对话"],"sub_categories":["大语言对话模型及数据"],"readme":"# **LaMMA-P: Generalizable Multi-Agent Long-Horizon Task Allocation and Planning with LM-Driven PDDL Planner**\n\nThis is the official repository for the LaMMA-P codebase. It includes instructions for configuring and running LaMMA-P on the MAT-THOR datasets in the AI2-THOR simulator. It is accepted as a conference paper by the IEEE International Conference on Robotics and Automation (ICRA), Atlanta, 2025.\n\n\n[Project Website](https://lamma-p.github.io/) | [Paper](https://arxiv.org/abs/2409.20560) | [Video](https://www.youtube.com/watch?v=1edDuJbk_uk)\n\n\u003cimg src=\"docs/motivation.png\" width=\"100%\"/\u003e\n\n**Abstract:** Language models (LMs) possess a strong capability to comprehend natural language, making them effective in translating human instructions into detailed plans for simple robot tasks. Nevertheless, it remains a significant challenge to handle long-horizon tasks, especially in subtask identification and allocation for cooperative heterogeneous robot teams. To address this issue, we propose a Language Model-Driven Multi-Agent PDDL Planner (LaMMA-P), a novel multi-agent task planning framework that achieves state-of-the-art performance on long-horizon tasks. LaMMA-P integrates the strengths of the LMs’ reasoning capability and the traditional heuristic search planner to achieve a high success rate and efficiency while demonstrating strong generalization across tasks. Additionally, we create MAT-THOR, a comprehensive benchmark that features household tasks with two different levels of complexity based on the AI2-THOR environment. The experimental results demonstrate that LaMMA-P achieves a 105% higher success rate and 36% higher efficiency than existing LM-based multi-agent planners.\n\n## Code Organization\nBelow are the details of various important directories \n- `resources/`: Contains robot definitions and PDDL domain files\n- `scripts/`: Main execution scripts adapted from [SMART-LLM](https://github.com/SMARTlab-Purdue/SMART-LLM)\n- `data/`: Test datasets and example tasks extended from [SMART-LLM](https://github.com/SMARTlab-Purdue/SMART-LLM)\n- `downward/`: Fast Downward planner from [Fast Downward](https://github.com/aibasel/downward/)\n\n## Datasets\nThe repository includes various commands and robots with different skill sets for heterogeneous robot tasks:\n\n- Test tasks: `data/final_test/`\n- Robot definitions: `resources/robots.py`\n- Floor plans: Refer to [AI2Thor Demo](https://ai2thor.allenai.org/demo) for layouts\n\n## Environment Setup\n### 1. Environment Setup\n\nCreate a conda environment (or virtualenv):\n```bash\nconda create -n lammap python==3.9\nconda activate lammap\n```\n\nInstall dependencies:\n```bash\npip install -r requirements.txt\n```\n\n### 2. Fast Downward Planner Setup\nThe project requires the [Fast Downward Planner](https://github.com/aibasel/downward/). Follow these steps to set it up:\n\n1. Clone the Fast Downward repository as a submodule:\n```bash\ngit submodule update --init --recursive\ncd downward\n```\n\n2. Build the planner:\n```bash\n./build.py\n```\n\n3. Verify the installation:\n```bash\n./fast-downward.py --help\n```\n\n### 3. OpenAI API Setup\nThe code relies on OpenAI's API for LLM functionality. To set this up:\n\n1. Create an API Key at https://platform.openai.com/\n2. Create a file named `api_key.txt` in the root folder\n3. Paste your OpenAI API Key in the file\n\n## Quickstart\n\n### 1. Generate PDDL Plans\nTo generate PDDL plans for tasks in AI2Thor floor plans, run:\n```bash\npython scripts/pddlrun_llmseparate.py --floor-plan \u003cfloor_plan_no\u003e\n```\n\nAdditional parameters:\n- `--gpt-version`: Choose between 'gpt-3.5-turbo', 'gpt-4o', 'gpt-3.5-turbo-16k' (default: 'gpt-4o')\n- `--prompt-decompse-set`: Set decomposition prompt set (default: 'pddl_train_task_decomposesep')\n- `--prompt-allocation-set`: Set allocation prompt set (default: 'pddl_train_task_allocationsep')\n\nThe script will:\n1. Decompose the high-level task into subtasks\n2. Generate PDDL problem files for each subtask\n3. Run the Fast Downward planner on each subtask\n4. Combine the solutions into a complete plan\n\nOutput files are stored in the `logs` directory, organized by timestamp and task name.\n\n### 2. Execute Plans in AI2Thor\nTo execute the generated plans in the AI2Thor environment:\nConvert the target plan into code\n```bash\npython plantocode.py --logs-dir ./logs --validate-code\n\n```\nthen, \n```bash\npython scripts/execute_plan.py --command \u003clog_folder_name\u003e\n```\nReplace `\u003clog_folder_name\u003e` with the specific folder name in the `logs` directory containing your generated plan.\n\n## Citation\nIf you find this work useful for your research, please consider citing:\n```bibtex\n@inproceedings{zhang2025lamma,\n  title={LaMMA-P: Generalizable Multi-Agent Long-Horizon Task Allocation and Planning with LM-Driven PDDL Planner},\n  author={Zhang, Xiaopan and Qin, Hao and Wang, Fuquan and Dong, Yue and Li, Jiachen},\n  booktitle={2025 IEEE International Conference on Robotics and Automation (ICRA)},\n  year={2025},\n  organization={IEEE}\n}\n```\n\n## Acknowledgement\n\nWe sincerely thank the researchers and developers for [SMART-LLM](https://github.com/SMARTlab-Purdue/SMART-LLM), [AI2THOR](https://github.com/allenai/ai2thor), and [Fast Downward](https://github.com/aibasel/downward/) for their amazing work.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftasl-lab%2FLaMMA-P","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ftasl-lab%2FLaMMA-P","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ftasl-lab%2FLaMMA-P/lists"}