需求侧响应的算力网络调度策略研究

Research on demand side response computing power network scheduling strategy

本文介绍了一种考虑需求侧响应的算力网络调度策略,包括算力系统调度方法、算力供给侧和需求侧的匹配、综合算力系统调度方法以及未来研究方向。通过将需求分类为可转移需求、可消减需求、可中断需求和不可消减需求,采用分时计价策略和可中断需求策略进行需求侧算力的调整,并调整各分布式数据中心设备算力的供给和边缘节点设备的接入,在满足综合算力系统的各约束前提下获得最低的系统总成本。该调度方法可以为算力网络系统的并网和智能调度提供一种新的思路。

This article mainly introduces a computing power network scheduling strategy that considers demand side response,including computing power system scheduling methods,problems with computing power supply and demand sides,comprehensive computing power system scheduling methods,and future research directions.By classifying requirements into transferable,reducible,interruptible and irreducible,using time-sharing pricing strategies and interruptible demand strategies to adjust demand side computing power,and adjusting the output of computing power for distributed data center equipment and the access of edge node equipment,the lowest total system cost is obtained while meeting the constraints of the comprehensive computing power system,This scheduling method can provide a new approach for the intelligent scheduling of computing power network systems.Finally,the article provides a brief introduction to the research direction of future intelligent scheduling in computing power networks.