Dynamic Programming‐Driven Multi‐Stage Decision‐Making Model for Zero‐Carbon Data Centers

Achieving lower carbon emission is important for data centers, and the “Channeling Computing Resources from the East to the West” (CCREW) project is proposed in China to develop zero‐carbon data centers, but the temporal dependency of renewable energy sources (RES) supply and demand remains a challenge. A multi‐stage decision‐making model based on dynamic programming is proposed in this paper to forecast available time windows of RES and schedule delay‐tolerant tasks to time windows. Simulation shows the decision‐making model can achieve a 94.83% task completion rate and 62.43% RES utilization.