In the chapter, we examine the possibility to use service dataflow extracted from logs of events of various elements for the analysis of the security of cyber-physical systems. The urgency of security providing for such complexes and their separate components is stipulated by a considerable number of security incidents as the violation of their faultless continuous functioning can lead to the damage of to people’s life and health, to the disruption of the functioning of the subsistence means of the population and other technogenic catastrophes. A normalization procedure of service dataflow is offered, allowing to generate groups of objects and properties and to ensure the process is such a way that semantically similar data are allocated in the same taxonomic fields. Besides, we examined the possibility to use the correlation of events in the conditions of uncertainty of infrastructure to define the hierarchy of asset types and the most significant objects. The approach is based on the detection of connections between events of security and the estimation of the dependence of their relative force on the distribution of events in time, based on the calculation and use of correlation coefficients. Such an approach allows us to reveal the main types of objects of the infrastructure, their characteristics, and hierarchy by analyzing empirical data, as well as to compare the criticality of objects based on static and dynamic indexes. We present the results of the experiment according to which the application of the given procedure allows us to lower information loss during the transformation of models of identified objects. The implementation of the studied approach to the correlation of events has some limitations in the case with a heterogeneous infrastructure; however, it can be used to develop a short-term pattern of interaction object with a homogeneous infrastructure in cyber-physical systems. Also, it is shown that automation of the detection of elements and the hierarchy of connections between them as well as the estimation of their criticality allows us to more precisely and in more detail develop a model of risks for the security analysis.

Generalization of the Lorden’s inequality is an excellent toolfor obtaining strong upper bounds for the convergence rate for variouscomplicated stochastic models. This paper demonstrates a method forobtaining such bounds for some generalization of the Markov modulatedPoisson process (MMPP). The proposed method can be applied in thereliability and queuing theory

Software verification and validation is the principal and most common type of activity for ensuring the quality of software systems, including those with a rising risk of the operation, throughout the entire life cycle. The software for systems important for nuclear power-plant (NPP safety) is subject to increased requirements for reliability and stability of operation, and for cybersecurity. The complex structure of automated process control systems of NPP and a large number of cybersecurity management tools make it difficult to verify and use cybersecurity management tools at all stages of the object’s life cycle from design to operation. The paper proposes a methodology for verifying and validating software for systems important to NPP safety, based on a high-qualitative cybersecurity management model to ensure the required level of cybersecurity. The methodology allows it to apply the concept of adaptive testing and simulate making and eliminating errors at each stage of software development. This methodology is using in the development of top-level systems of NPP APCS.

The applicability of the data on the mean age of death (MAD) of people presented on Wikipedia to determine the survival patterns with respect to professional occupations is considered. The results of previous studies are described. Features of these data related to censorship and the selective nature of their collection are discussed. The proposed approach to account for these features makes it possible to clarify the obtained patterns and to expand the time range of the study.

We consider the nonparametric estimation of the extremal index of stochastic processes. The discrepancy method that was proposed by the author as a data-driven smoothing tool for probability density function estimation is extended to find a threshold parameter u for an extremal index estimator in case of heavy-tailed distributions. To this end, the discrepancy statistics are based on the von Mises– Smirnov statistic and the k largest order statistics instead of an entire sample. The asymptotic chi-squared distribution of the discrepancy measure is derived. Its quan8 tiles may be used as discrepancy values. An algorithm to select u for an estimator of the extremal index is proposed. The accuracy of the discrepancy method is checked by a simulation study.