SECURITY OF COMPUTER NETWORKS IN CONDITIONS OF UNCERTAINTY

Abstract

In the modern era of digitalization, the issue of information protection in various types of computer networks is becoming increasingly relevant. The constant emergence of new threats, the dynamic nature of information flows, and the unpredictable behavior of attackers create an environment with a high level of uncertainty. This complicates the application of traditional methods of risk assessment and security system design. This article examines theoretical and practical approaches to ensuring network security under uncertainty. Methods of risk modeling using probabilistic and fuzzy models are analyzed, and the role of adaptive algorithms and machine learning in threat detection is outlined. A conceptual framework is proposed for building comprehensive protection systems capable of self-learning and adapting to a changing environment. Special emphasis is placed on the use of mathematical models capable of accounting for incomplete data and fuzziness in defining security parameters. The combination of probabilistic analysis, fuzzy logic methods, and machine learning algorithms enables the development of adaptive systems that can respond promptly to emerging threats. Risk modeling under uncertainty allows for a quantitative assessment of the effectiveness of network protection measures. The fuzzy approach and fuzzy variables make it possible to consider incomplete or imprecise information about potential threats. Adaptive models prove to be more effective compared to static ones, reducing the maximum risk of attacks. The use of scenario modeling (DoS, MITM) helps identify critical network points and plan optimal protection measures. The work contributes to the formation of theoretical and practical foundations for ensuring network security under uncertainty and substantiates methods that enhance the resilience of information systems to unpredictable attacks and reduce the risks of data loss.