| 引用本文: | 屈径,刘文泽,蔡泽祥,岑伯维,胡凯强,刘媛媛.配电物联网云边协同系统计算资源双目标优化配置方法[J].电力自动化设备,2024,44(5):112-119. |
| QU Jing,LIU Wenze,CAI Zexiang,CEN Bowei,HU Kaiqiang,LIU Yuanyuan.Dual-objective optimal configuration method for computing resources of cloud-edge collaborative system in power distribution Internet of Things[J].Electric Power Automation Equipment,2024,44(5):112-119. |
|
| 摘要: |
| 云边协同系统能够支持计算资源弹性扩展，适应配电物联网技术的需求。针对云边协同系统的计算资源配置问题，刻画了云边协同系统的计算资源，建立了计算业务模型，提出了批量计算业务的业务超时系数以对系统效果进行定量评价。基于上述模型，以最小化云边协同系统开销和最小化业务超时系数为目标，建立了考虑云边互动的计算资源优化配置双目标规划模型，并采用改进差分进化算法得到帕累托前沿。基于仿真算例讨论了不同双目标处理方法、通信质量、并发业务规模等对所提计算资源配置方法的影响。 |
| 关键词: 配电物联网 计算资源配置 云边协同系统 双目标规划 差分进化算法 |
| DOI：10.16081/j.epae.202311026 |
| 分类号: |
| 基金项目:广东省重点领域研发计划项目(2019B111109002) |
|
| Dual-objective optimal configuration method for computing resources of cloud-edge collaborative system in power distribution Internet of Things |
|
QU Jing, LIU Wenze, CAI Zexiang, CEN Bowei, HU Kaiqiang, LIU Yuanyuan
|
|
School of Electric Power, South China University of Technology, Guangzhou 510641, China
|
| Abstract: |
| The cloud-edge collaborative system can support the elastic expansion of computing resources and adapt to the needs of power distribution Internet of Things technology. Aiming at the computing resource configuration problem of cloud-edge collaborative system, the computing resources of the cloud-edge collabo-rative system are described, the computing business model is established, and the business timeout coefficient of batch computing business is proposed to evaluate the system effect quantitatively. Based on the above models, a dual-objective programming model for optimization configuration of computing resources considering cloud-edge interaction is established with the goals of minimizing the system overhead and minimizing the business timeout coefficient. The Pareto frontier is obtained by using the improved differential evolution algorithm. Based on the simulation examples, the influence of different dual-objective processing methods, communication quality and concurrent business size on the proposed computing resource configuration method is discussed. |
| Key words: power distribution Internet of Things configuration of computing resources cloud-edge collaborative system dual-objective programming differential evolution algorithm |
