SustainDC
收藏资源简介:
SustainDC是由惠普企业(惠普实验室)开发的一个用于可持续数据中心控制的多代理强化学习(MARL)环境集合。该数据集支持自定义数据中心配置和任务,如工作负载调度、冷却优化和辅助电池管理。数据集内容包括多样化的数据中心设计、位置、天气条件、电网碳强度和工作负载需求。创建过程中,SustainDC允许用户模拟大型数据中心的电气和热流体行为,并进行定制化设计。该数据集主要应用于优化数据中心操作,以减少碳足迹和提高能源效率,适用于AI研究和环境可持续性问题的解决。
SustainDC is a collection of multi-agent reinforcement learning (MARL) environments for sustainable data center control, developed by Hewlett Packard Enterprise (HPE Labs). This dataset supports customizable data center configurations and tasks, such as workload scheduling, cooling optimization, and auxiliary battery management. The dataset includes diverse data center designs, locations, weather conditions, grid carbon intensity, and workload demands. In its development, SustainDC enables users to simulate the electrical and thermofluid behaviors of large-scale data centers and conduct customized designs. This dataset is primarily used to optimize data center operations to reduce carbon footprints and improve energy efficiency, and is applicable to AI research and addressing environmental sustainability challenges.

- 1SustainDC -- Benchmarking for Sustainable Data Center Control惠普企业(惠普实验室) · 2024年



