Search papers, labs, and topics across Lattice.
This paper introduces PowerAtlas, an LLM-agent framework designed for coordinated electricity-computing co-scheduling in data centers, addressing the challenges posed by the increasing volatility of AI workloads. By integrating historical data, domain knowledge, and physical constraints, PowerAtlas generates schedules that comply with grid operational rules and service-level agreements (SLAs). Experimental validation with a provincial power utility in China shows that PowerAtlas consistently achieves feasible and cost-effective scheduling solutions across various LLM architectures.
PowerAtlas achieves optimal electricity-computing scheduling, ensuring compliance with grid rules while maximizing efficiency in volatile AI workloads.
The rapid growth of AI workloads is turning data centers into large-scale, volatile, yet spatiotemporally flexible grid loads, creating an urgent need for coordinated electricity-computing scheduling. Under stringent grid constraints, schedules from general-purpose large language models (LLMs) are often infeasible, causing line-flow violations and unserved load. We present PowerAtlas, an LLM-agent framework for electricity-computing co-scheduling that integrates historical instances, domain knowledge, and physical constraints to produce joint decisions satisfying both grid operational rules and the service-level agreements (SLAs) of computing tasks. Working with a provincial power utility in China, we built an experimental electricity-computing network and validated the decision loop on real data-center data; from de-identified operational data we further constructed ECBench, a benchmark of 2,000 scheduling instances with oracle-optimal solutions. Experiments across eleven LLMs demonstrate the effectiveness of PowerAtlas under realistic physical operating conditions, with consistent feasibility and cost gains across three open-weight backbones. Our code is publicly available at https://github.com/JAVA-Jiang/PowerAtlas.
