aditijc/snooker-testbed-phase4g-obsaug-v1

该数据集记录了在斯诺克测试平台上进行的强化学习实验的结果，具体为使用SAC算法在增强79维观察空间下的性能数据。实验旨在通过添加手工制作的特征（如最佳目标的正弦/余弦、标准化距离、对齐分数等）来改善策略对斯诺克几何的理解。数据集包含10行8列，记录了实验过程中的时间步、平均得分、最高得分、平均犯规率等关键指标。实验结果显示，观察增强提高了稳定性，但未能显著降低犯规率。

This dataset records the results of a reinforcement learning experiment conducted on a snooker testbed, specifically the performance data of the SAC algorithm in an augmented 79-dimensional observation space. The experiment aimed to improve the policys understanding of snooker geometry by adding hand-crafted features such as best-target sin/cos, normalized distances, alignment score, etc. The dataset contains 10 rows and 8 columns, documenting key metrics during the experiment such as timesteps, mean score, max score, and mean foul rate. The results show that observation augmentation improves stability but does not significantly reduce the foul rate plateau.