Carbon-Aware Compute--Power Scheduling for AI Data Centers with Microgrid Prosumer Operations
For operators of geographically distributed AI data centers, this work provides a tractable optimization framework to reduce carbon emissions while improving operational benefit, though it is an incremental extension of existing MILP approaches to a specific domain.
This paper proposes a mixed-integer linear programming framework for carbon-aware compute-power scheduling in AI data centers with microgrid prosumer capabilities, jointly optimizing training jobs, inference routing, local generation, and battery storage. Experiments show substantial improvements in operational benefit and emission reductions over baselines, with inference routing flexibility being a major value driver.
AI data centers are increasingly becoming tightly coupled compute--energy systems, where workload placement, cooling demand, electricity procurement, storage operation, and carbon emissions interact over time. This paper studies carbon-aware compute--power scheduling for geographically distributed AI data centers with microgrid prosumer capabilities. We propose a mixed-integer linear programming (MILP) framework that jointly schedules rigid training jobs, routes elastic inference workloads, dispatches local generation and battery storage, and manages bidirectional grid interaction under latency, continuity, power-balance, and carbon-budget constraints. The model captures two key features of emerging AI infrastructure: heterogeneous workload flexibility and site-level energy prosumer operation. Experiments on synthetic yet practically motivated instances show that the proposed joint MILP substantially improves total operational benefit over compute-only and energy-only baselines while reducing emissions. The results further indicate that inference-routing flexibility is a major source of value, battery storage provides useful temporal flexibility, and local-generation-rich settings are particularly favorable. The framework provides a tractable optimization abstraction for sustainable and grid-interactive AI data centers.