A Study on the Flexible Scheduling Strategy for the Computing Power Load of Data Centers Based on Multi-algorithm Integration

Under the dual drive of the East-to-West Computing Resource Allocation and Dual Carbon strategies, the contradiction between the rigid growth of computing power load in data centers and the pressure of new energy consumption in the power grid has become increasingly prominent. This paper proposes and practices a complete theoretical and technical system for the collaborative optimization of computing and power, aiming to transform data centers from traditional rigid loads into flexible resources that support the construction of a new power system. This paper mainly makes the following three innovations: First, a closed-loop system engineering framework of monitoring dispatching evaluation has been constructed, achieving fullprocess management from panoramic monitoring of electricity and carbon, intelligent integrated dispatching to real-time feedback of effects. Secondly, a Power Computing Power Intraday Fusion Scheduling Optimization Model integrating multiple algorithms was proposed. This model adopts dual time scales of a minute level and a 15-minute level, aiming to minimize the total cost. It innovatively combines greedy algorithms, simulated annealing, and task exchange strategies to find an efficient solution, guiding the computing power load curve to actively follow the photovoltaic output curve. Thirdly, by integrating multimodal carbon data-driven and AI mechanismdriven hybrid modeling, a carbon emission knowledge graph and causal reasoning model have been constructed, enhancing the interdependence and strategy adaptation capabilities of carbon management. An empirical study taking the State Grid Jiangxi Electric Power Data Center as the object shows that this system can achieve cost savings in operation and effectively improve the utilization rate of clean energy. This research provides a replicable and scalable systematic solution for solving the problem of spatiotemporal mismatch of computer resources, which is of great theoretical and practical significance for promoting the green and low-carbon coordinated development of energy and digital infrastructure.