DataCenterGym: A Physics-Grounded Simulator for Multi-Objective Data Center Scheduling

2026-04-16

AI Summary

This study addresses the neglect of coupled computational, thermal, and energy dynamics in scheduling heterogeneous workloads across geographically distributed data centers. To this end, the authors develop DataCenterGym, a physics-informed simulation environment that, for the first time, integrates a full thermo-electrical-computational coupling mechanism into a scheduling framework. It explicitly models building thermal dynamics, localized HVAC behavior, temperature-dependent service degradation, and task queuing, while maintaining compatibility with the Gymnasium interface. Building upon this platform, they propose a hierarchical model predictive control (H-MPC) algorithm that explicitly co-optimizes thermal and electrical dynamics. Experimental results demonstrate that H-MPC significantly outperforms baseline methods under both typical and load-sensitive scenarios, achieving a superior trade-off among energy efficiency, cooling costs, and quality of service.

Technology Category

• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Machine Learning: Hardware-aware ML
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Modern datacenters schedule heterogeneous workloads across geo-distributed sites with diverse compute capacities, electricity prices, and thermal conditions. Compute utilization, heat generation, cooling demand, and energy consumption are tightly coupled, yet most existing schedulers abstract these effects and treat them independently. We present \textit{DataCenterGym}, a physics-grounded simulation environment for job scheduling in geo-distributed data centers, designed as a reusable testbed for future research. The simulator integrates compute queueing, building thermal dynamics, localized HVAC behavior, and temperature-dependent service degradation within a Gymnasium-compatible interface. We also develop a Hierarchical Model Predictive Control (H-MPC) scheduling algorithm that performs distributed job placement while explicitly accounting for thermal and power dynamics. Through experiments on nominal operation and workload sensitivity, we demonstrate how H-MPC improves scheduling performance relative to baseline schedulers.

Problem

• data center scheduling
• thermal dynamics
• energy consumption
• geo-distributed systems
• workload heterogeneity

Innovation

• physics-grounded simulation
• geo-distributed data centers
• thermal-aware scheduling
• hierarchical model predictive control
• multi-objective optimization