综合智慧能源2026，Vol.48Issue(8)：82-95,14.DOI:10.3969/j.issn.2097-0706.2026.08.007
考虑低碳激励的运营商-数据中心群主从博弈调度策略
A Stackelberg game-based scheduling strategy for operator-data center clusters considering low-carbon incentives
摘要
Abstract
Driven by the dual strategic objectives of achieving the"dual carbon"goals and promoting high-quality development of the digital economy,data center clusters continue to expand in scale.The intensive energy consumption of data centers has increasingly intensified conflicts involving operating costs,carbon emissions,and benefit allocation among multiple stakeholders,and existing operation modes can hardly balance economic efficiency and low-carbon performance.To address these issues,a Stackelberg leader-follower scheduling method for an integrated energy operator and data center cluster considering low-carbon incentives was proposed.A coupled cooling-heating-electricity model incorporating gas-fired power generation units,gas boilers,absorption chillers,heat pumps,wind power,and photovoltaic generation was established to characterize multi-energy conversion,renewable energy utilization,and operational relationships within data centers.Then,a Stackelberg game model was constructed with the integrated energy operator as the leader and the data center cluster as the follower.The operator determined electricity prices,gas prices,and low-carbon incentive strategies based on upstream energy prices and low-carbon operation requirements.Each data center independently solved for its optimal response subject to energy balance and equipment operation constraints,and autonomously adjusted its energy consumption patterns.On this basis,an external low-carbon compensation mechanism based on actual effective emission reductions was introduced to alleviate the profit pressure faced by the operator from providing low-carbon incentives.Furthermore,maximizing operator profit,minimizing the net operating cost of the data center cluster,and minimizing system carbon emissions were formulated as the three upper-level optimization objectives.The multi-objective evolutionary algorithm based on decomposition was employed to search for a Pareto non-dominated strategy set subject to the optimal-response constraints of the data centers,from which a compromise scheduling solution was selected.Case study results demonstrated that the proposed method could effectively coordinate operator profit,data center operating costs,and system carbon emission reduction objectives.Compared with the baseline scenario,the proposed game mechanism reduced the net operating cost of the data center cluster by 3.49%and system carbon emissions by 1.96%.Compared with the scenario considering only internal low-carbon incentives,the introduction of external low-carbon compensation increased operator profit by 7.41%.The findings indicate that the proposed method can balance the economic efficiency and low-carbon performance of multiple stakeholders while maintaining the autonomous response capability of data centers,thereby providing a reference for integrated energy optimization scheduling and low-carbon coordinated operation of data center clusters.关键词
数据中心/综合能源运营商/Stackelberg博弈/基于分解的多目标进化算法/低碳激励/算电协同Key words
data center/integrated energy operator/Stackelberg game/multi-objective evolutionary algorithm based on decomposition/low-carbon incentive/computing-electricity coordination分类
能源科技引用本文复制引用
龚钢军,宋得山,陈昭赫,柴凯天,杨佳轩..考虑低碳激励的运营商-数据中心群主从博弈调度策略[J].综合智慧能源,2026,48(8):82-95,14.基金项目
国家电网有限公司科技项目(2025YF-120) （2025YF-120）
国家自然科学基金项目(52477095)Science and Technology Project of State Grid Corporation of China(2025YF-120) （52477095）
National Natural Science Foundation of China(52477095) （52477095）
