CHiRP

该数据集名为CHiRP，在具有挑战性的连续强化学习环境中进行了评估，涵盖了多种不同的领域，重点关注各种环境的随机版本。该数据集的编制旨在评估CHiRP方法在连续强化学习中的有效性，特别关注样本效率和任务分解等关键维度。其规模相较于之前的调查研究有所扩大，任务重点在于具有层次分解的连续强化学习。

This dataset, named CHiRP, was evaluated across challenging continuous reinforcement learning environments spanning diverse domains, with a primary focus on stochastic variants of these environments. The dataset was compiled to assess the effectiveness of the CHiRP method in continuous reinforcement learning, with particular emphasis on key dimensions such as sample efficiency and task decomposition. Its scale is larger than that of prior survey-based studies, and its tasks focus on continuous reinforcement learning with hierarchical decomposition.