| 引用本文: | 郭 伟,范文奕,杨书强,等.面向分布式一致性算法的通信网络优化设计[J].电力系统保护与控制,2022,50(23):151-160.[点击复制] |
| GUO Wei,FAN Wenyi,YANG Shuqiang,et al.Optimal design of a communication network for a distributed consensus algorithm[J].Power System Protection and Control,2022,50(23):151-160[点击复制] |
|
| 摘要: |
| 通信网络是实现分布式控制的基础设施。针对分布式一致性算法以及未来智能微电网的应用需求，提出了一种兼顾动态性、延迟鲁棒性和经济性的通信网络优化设计方法。首先，根据代数图论相关知识建立通信网络与相应矩阵的联系。其次，由不同矩阵定义了与通信网络相关的3个性能指标，并利用奈奎斯特稳定判据推导出最大通信延迟时间? 与拉普拉斯矩阵L特征值之间的关系。最后，由代数连通度相关定理，给出一种边数递减循环多目标优化方法。每次循环建立包含3个指标的多目标优化模型，并采用NSGA-II算法求解该边数下的满意解。重复上述过程直至网络不连通，根据网络的动态性和延迟鲁棒性选出所有满意解中的最终优化网络。仿真算例验证了所提优化方法的可行性和有效性。 |
| 关键词: 通信 分布式 拉普拉斯矩阵 代数连通度 延迟 优化 |
| DOI：DOI: 10.19783/j.cnki.pspc.221155 |
| 投稿时间：2022-07-21修订日期：2022-10-25 |
| 基金项目:河北省重点研发计划项目资助(21314302D)；国家电网有限公司科技项目(5400-202155371A-0-0-00) |
|
| Optimal design of a communication network for a distributed consensus algorithm |
| GUO Wei,FAN Wenyi,YANG Shuqiang,AN Jiakun,HE Chunguang,WANG Tao,JING Tianjun |
| (1. State Grid Hebei Economic Research Institute, Shijiazhuang 050000, China; 2. College of Information and Electrical
Engineering, China Agricultural University, Beijing 100083, China) |
| Abstract: |
| Communication networks are the infrastructure for distributed control. For a distributed consensus algorithm and future smart microgrid application requirements, a communication network optimization design method that takes into account dynamics, delay robustness, and economy is proposed. First, the relation between the communication network and the corresponding matrix is established by algebraic graph theory. Then, three performance indices related to the communication network are defined by different matrices, and the relationship between the maximum communication delay time ? and the Laplacian matrix L eigenvalues is deduced using the Nyquist stability criterion. Finally, using an algebraic connectivity related theorem, an edge decrement cycle network optimization process is proposed. A multi-objective optimization model including three indices is established in each cycle, and a solution is achieved using the NSGA-II algorithm for the number of edges. The process is repeated until the network is disconnected, and the final optimization network is selected from all the satisfactory solutions according to the network dynamics and delay robustness. Simulation examples verify the feasibility and effectiveness of the proposed optimization method.
This work is supported by the Key Research and Development Program of Hebei Province (No. 21314302D). |
| Key words: communication distributed Laplacian matrix algebraic connectivity delay optimization |
