Open Access

ARTICLE

Multi-Objective Optimization of Data Center Computing-Electricity-Cooling-Heating Coupled System Considering Economy, Flexibility, and User Satisfaction

Dongjunming Yang¹, Junqiu Fan^1,*, Qingsheng Li¹, Yu Zhang¹, Renren Du¹, Jiangjiang Wang²
1 Guizhou Power Grid Co., Ltd., Guiyang, China
2 Yanzhao Electric Power Laboratory, North China Electric Power University, Baoding, China
* Corresponding Author: Junqiu Fan. Email:
(This article belongs to the Special Issue: Clean Energy and Low-Grade Energy Utilization: Material, Component, and System Innovation)

Energy Engineering https://doi.org/10.32604/ee.2026.087225

Received 12 June 2026; Accepted 14 July 2026; Published online 27 July 2026

Abstract

Data centers (DC) have significant potential to provide grid-side flexibility, but the coordinated contribution of computing workload shifting, hybrid cooling, and waste heat recovery has not been sufficiently quantified in existing studies. To address this gap, this study proposes a holistic framework for optimizing the integrated computing-electricity-cooling-heating nexus of a DC within a power grid, integrating computing tasks, power and cooling supplies, and waste heat recovery. A multi-objective, bi-level stochastic optimization model is developed to co-optimize strategic equipment capacity and operational dispatch, simultaneously targeting economic efficiency, operational flexibility, and computational user satisfaction. The framework quantifies cross-domain flexibility through a novel aggregated indicator encompassing electrical, cooling, and computational adjustments. A scenario-based approach handles uncertainties in renewable generation, workload, and load. A comprehensive case study validates the framework, involving the construction of representative operational scenarios, Pareto frontier analysis to reveal trade-offs, and comparative analysis of distinct operational strategies (cost-, flexibility-, and satisfaction-oriented). The results confirm inherent conflicts among the objectives and demonstrate that the proposed coordinated optimization can effectively reshape the DC’s power profile through computational load shifting. Achieving high service quality incurs a measurable cost premium, while a flexibility-oriented strategy can enhance the system’s adjustment capacity by over 27% compared to a cost-minimal approach, albeit with increased reliance on grid interaction. The proposed integration reduces the economic cost by approximately 10.4% compared with the case without workload shifting and by approximately 3.2% compared with the electric-chiller-only cooling case, while maintaining an acceptable customer satisfaction level.

---

Keywords

Computing-electricity coordination; data center (DC); hybrid cooling; multi-objective optimization; operational flexibility; waste heat recovery

---