Vol 26, No 7 (2020)

A non-Markov model of a random process is considered, in which a simple stream arrives at the input of the system, and the duration between states is distributed according to an arbitrary law. The equation of the connection of the semi-Markov process with fuzzy and clear states is compiled. Various particular cases of such a connection are given, when the given triangular membership functions of fuzzy states are modified into their unit analogues, and the kernels of the integral equations are represented by an exponential distribution. A description of the recovery process as a special case of the semi-Markov process and the appearance of a fuzzy integral equation is presented.

The piper considers the task of organizing a rational interaction between a geographic information service (GIS) and a user who analyzes spatial data. The purpose of the analysis is to develop a solution to an applied problem in conditions of incompleteness and uncertainty of information about both the final goal and the solution procedure. The work considers an interactive visual analysis of a space region, where the GIS delivers data from cartographic sources and the user compiles and analyzes the cartographic representation of the workspace in real-time. The task of guiding the analysis process to maintain its intelligent orientation is formulated. The meaning of the analysis process is associated with achieving situation awareness of the user, which will ensure the adoption of higher qualitydecisions. A control method based on the selection of an adequate context from the set of contexts described in the GIS has proposed. A two-component context model has proposed in the form of a kernel and a set of valid kernel transformations preserving the meaning of the context. The intelligent orientation of the analysis has stabilized using the proposed presentation. The problem of determining the optimal number of GIS contexts has considered which minimizes the complexity of the analysis and relation has obtained for its calculation. The problem of maximizing the utility of cartographic images generated by GIS at the user’s request has formulated. A representation of the analysis workspace by skeleton and environment has proposed, which has used to ensure maximum utility. The results of an experimental study of the layout of an intelligent system for managing the analysis are presented too.

In this article were discussed the using of the genetic algorithm of differential evolution to determine the orientation and position of an object relative to an optical sensor by means a series of frames. In this case, commercial optical cameras are used as measuring means. A statistical study of the genetic algorithm of differential evolution was conducted. The study was conducted on the model problem of separation of two objects. Validation of estimates is carried out using the nonlinear least squares method. The paper presents the dependencies of errors in determining the relative orientation and relative position obtained using two algorithms. The study shows that the genetic algorithm of differential evolution can be used in various problems of processing information from optical sensors.

Description of the hierarchical model of the characteristics of the economic system is the goal. The model is made by taxonomic principle. The hierarchy of models on the example of indicators of profit, competitiveness and customer loyalty is shown. A stratified description of the object of study is the result. To study the model, simplicial analysis was used. Simulation experiments conducted. A comparison of the results before and after simplicial analysis is done.

The problem of determining the information security state of an unmanned vehicle by an external monitoring system was considered. The effectiveness of solving this problem is estimated by the overall accuracy of the classification of the state of the object. A method based on the analysis and classification of spectral data recorded by a digital three-axis accelerometer mounted on board a vehicle was proposed. An experiment aimed at studying the proposed method was described. As a result of the experiment, the fundamental possibility of using accelerometer in information security monitoring systems is shown. The overall accuracy of the selected classifier based on the k-NN algorithm was 0.82.