DMH Instances
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/HcPlu/DMH-GYM
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了从不同分布中抽取的训练实例（DMH-01至DMH-08）以及通过给训练实例添加噪声生成的测试实例（DMH-09至DMH-16）。该数据集用于评估ACERL方法在动态物料处理任务中的性能。此外，该数据集还用于测试包括ACERL、SAC和PPO在内的多种强化学习方法的性能，并包含了诸如完工时间、延迟以及约束满足率等指标。规模上，该数据集包含了8个训练实例和8个测试实例，其任务针对的是动态物料处理。
This dataset contains training instances sampled from distinct distributions (DMH-01 to DMH-08) and test instances generated by adding noise to the training instances (DMH-09 to DMH-16). It is used to evaluate the performance of the ACERL method on dynamic material handling tasks. Additionally, this dataset is also employed to test the performance of multiple reinforcement learning methods including ACERL, SAC and PPO, and includes metrics such as makespan, delay and constraint satisfaction rate. In terms of scale, this dataset comprises 8 training instances and 8 test instances, all targeting dynamic material handling tasks.
提供机构：
Publicly available through the DMH-GYM simulator
