IPC-2023 Total-Order Track
收藏资源简介:
IPC-2023 Total-Order Track数据集由2023年国际规划竞赛(IPC-2023)的分层任务网络(HTN)规划赛道提供,包含22个领域,每个领域有数十个问题。该数据集旨在为分层规划(HP)领域的研究提供标准化的评估基准,特别是用于评估基于大语言模型(LLM)的规划方法的性能。数据集的内容涵盖了从问题定义到计划生成的多个阶段,数据来源为IPC-2023的官方竞赛数据。该数据集的创建过程基于竞赛中的实际规划问题,旨在解决分层规划领域中LLM集成方法的性能评估问题,并为未来的研究提供基础参考。
The IPC-2023 Total-Order Track dataset is provided by the Hierarchical Task Network (HTN) Planning Track of the 2023 International Planning Competition (IPC-2023), consisting of 22 domains, with dozens of problems in each domain. This dataset is designed to serve as a standardized evaluation benchmark for the Hierarchical Planning (HP) field, particularly for assessing the performance of planning methods based on Large Language Models (LLM). The content of the dataset covers multiple stages from problem definition to plan generation, and its data source is the official competition data of IPC-2023. The creation process of this dataset is based on actual planning problems from the competition, aiming to address the performance evaluation issues of LLM integration methods in the hierarchical planning field, and to provide a foundational reference for future research.




