Identification of key nodes in ship information flow network based on entropy weight method and FMEA

As a typical cyber-physical system (CPS), ship system, with the continuous improvement of its informatization degree, CPS information security incidents occur frequently. Most of the existing key node identification methods for complex networks are aimed at power system networks, and the evaluation...

Full description

Bibliographic Details
Main Author: LIU Xingyun, WANG Quanhu, WANG Jun, SHEN Naijun
Format: Article
Language:zho
Published: Editorial Office of Command Control and Simulation 2023-08-01
Series:Zhihui kongzhi yu fangzhen
Subjects:
Online Access:https://www.zhkzyfz.cn/fileup/1673-3819/PDF/1673-3819(2023)04-0087-05.pdf
Description
Summary:As a typical cyber-physical system (CPS), ship system, with the continuous improvement of its informatization degree, CPS information security incidents occur frequently. Most of the existing key node identification methods for complex networks are aimed at power system networks, and the evaluation indicators based on these networks are difficult to apply to CPS networks outside the power system. Aiming at the shortcomings of the existing key node identification methods in CPS, this paper proposes an information network key node identification method based on entropy weight method and FMEA. The concept of effective influence range is added to the indicator characterization of failure mode and effects analysis (FMEA), and the entropy weight method is used to objectively weight the indicators in FMEA. Combined with the risk sequence number in the FMEA with strong subjectivity, the three evaluation indicators in the risk sequence number are used as the indicators of the decision matrix in the entropy weight method. Finally, the correctness and feasibility of the method are verified by the case of ship support mission information flow node network. The analysis results show that the comprehensive weighted ranking of the three indicators based on the entropy weight method is more reasonable than a single indicator.
ISSN:1673-3819