QDax

该数据集是一个专注于机器人控制任务的深度神经进化在强化学习中的基准测试套件，特别关注全方位蚂蚁任务。它包含了多种质量多样性指标，用于算法评估、鲁棒性分析，以及不同质量多样性算法之间的比较。该数据集的任务是利用神经进化在强化学习中实现机器人控制。This dataset is a benchmark suite for deep neuroevolution in reinforcement learning, focusing on robotic control tasks with a particular emphasis on the omnidirectional ant task. It includes multiple quality-diversity metrics for algorithm evaluation, robustness analysis, and comparative assessments between different quality-diversity algorithms. The core task of this dataset is to achieve robotic control via neuroevolution within reinforcement learning frameworks.