SCADA安全因素神经元的云推理机研究与仿真

PDF下载HTML

Title:

Study and simulation of the SCADA security factors neuron’scloud inference engine

作者:

熊柳¹, 曹谢东¹, 李杰¹, 杨力², 刘增良³

1. 西南石油大学 电气信息学院, 四川 成都 610500;
2. 西南石油大学 计算机科学学院, 四川 成都 610500;
3. 中国人民解放军国防大学 信息作战研究所, 北京 100091

Author(s):

XIONG Liu¹, CAO Xiedong¹, LI Jie¹, YANG Li², LIU Zengliang³

1. School of Electrical Information Engineering, Southwest Petroleum University, Chengdu 610500, China;
2. School of Computer Science, Southwest Petroleum University, Chengdu 610500, China;
3. Institute of Information Operation, University of National Defense, Beijing 100091, China

关键词:

SCADA信息安全; 因素空间; 因素神经元; 云模型:MATLAB仿真

Keywords:

SCADA information security; factor space; factor neuron; cloud model; MATLAB simulation

分类号:

TP18

DOI:

10.11992/tis.201509020

摘要:

为了解决SCADA系统信息安全的问题，本文提出了一种基于因素神经网络的主动防御方法。将SCADA系统信息安全的影响因素映射到因素空间坐标中，然后利用知识因素的因素神经元表示方法，通过云模型推理机实现了语言值表示的模糊概念到定量数据的转换，并通过云模型多规则多条件发生器进行规则推理，最后根据得到的期望值又可以转换为定性语言值，这样就实现了对未知恶意程序行为操作可能性的预测。本文着重于利用基于云模型的多条件多规则发生器实现推理，通过MATLAB进行算法设计和仿真，为油气SCADA系统信息安全防御的解决方法提供了一种思路。

Abstract:

Over recent years, with the amount of incidents involving industrial control information systems, the safety of these system has been given increased importance across the Globe, and several technical measures have been implented to improve this. This paper proposed an active defense method based on a factor neural network in reference to SCADA information security. The different aspects of SCADA information security were mapped to factor coordinates, and the factor neuron method for knowledge factors was then used to transform this from a fuzzy concept (represented by language) to quantitative data through a cloud inference engine. Inference was then conducted through generated multi conditions and multi rules based on a cloud model, so that it could then be transformed into a qualitative language (e.g.‘more likely’based on Ex) to be able to forecast the consequences of unknown malicious programs. This paper focus on using a generator with multiple conditions and rules based on a cloud model to achieve inference, thus providing an idea of the oil and gas SCADA information security’s defense response using algorithmic design and MATLAB-based simulations.