A mathematical model and a method for the problem of optimization of aggregation and of sequential- parallel execution modes of intersecting operation sets are proposed. The proposed method is based on the two-level decomposition scheme. At the top level the variant of aggregation for groups of operations is selected, and at the lower level the execution modes of operations are optimized for a fixed version of aggregation.

An analytical solution of the boundary-value problem of penetration of electromagnetic field of magnetic dipole through the planar lens from the focusing materials is suggested. It is demonstrated, that for the lenses from focusing materials two focuses are formed: first is over the screen and the second is in front of the screen in the source area. The field of focus in front of the lens is equal to the sum of the fields of magnetic and electric dipoles.

An approach to solvе a problem of determining optimal dynamic prices for hotel rooms is suggested. It includes selection of input parameters for the succeeding mathematical analysis, disaggregation of the demand into several categories, demand forecasting, simulation of demand- price relations, and a mathematical programming model for price optimization.

The paper discusses the ways of increasing the efficiency of production of hydraulic drives by using the information support of product life cycle. The article describes examples of the use of software systems for analyzes at different stages of the life cycle of hydraulic drives.

With this study, a method for quantitative description of biomedical images based on splitting the target image into superpixels followed by categorization using precalculated superpixel dictionaries is proposed. The method has been tested on the tasks of recognition of biomedical images of three types: lung CT images, histology images of ovary and thyroid tissues. The results of the experiments performed suggest that the method proposed may provide recognition performance comparable or better than when using conventional methods of texture description.

The principles of fast parallel iterative algorithms based on the use of graphics accelerators and OpenGL library are considered in the paper. The proposed approach provides simultaneous minimization of the residuals of the desired solution and total variation of the reconstructed three- dimensional image. The number of necessary input data, i. e. conical X-ray projections, can be reduced several times. It means in a corresponding number of times the possibility to reduce radiation exposure to the patient. At the same time maintain the necessary contrast and spatial resolution of threedimensional image of the patient. Heuristic iterative algorithm can be used as an alternative to the well-known three-dimensional Feldkamp algorithm.

We analyze the existing DS from the point of security and construct a two-level hierarchy of models. Such approach allows us to separate the abstraction (architecture) level and the concrete (component) level of ISS. The core set of methods, i. e. authentication and key exchange protocols, corresponds to the abstraction level and is defined as security infrastructure (SI). The final security parameters optimization and additional mechanisms such as authorization, routing and data auditing of the protection mechanisms are configured on the component level of the DS. In addition, we outline the systematic step-by-step ISS configuration method.

The article presents the analysis of Grid-systems information security aspects, Also, risk assessment methods, using public databases, expert systems and statistical techniques are proposed.

The paper presents a method of construction of genetic data clusters (functional modules) using the randomized matrices. To build the functional modules the selection and analysis of the eigenvalues of the gene profiles correlation matrix is performed. The principal components, corresponding to the eigenvalues, which are significantly different from those obtained for the randomly generated correlation matrix, are used for the analysis. Each selected principal component forms gene cluster. In a comparative experiment with the analogs the proposed method shows the advantage in allocating statistically significant different-sized clusters, the ability to filter non- informative genes and to extract the biologically interpretable functional modules matching the real data structure.

Analysis of whole-genome sequences often leads to problems of large dimensionality where the number of parameters exceeds the number of available observations. We offer methodology of genome-wide association study and investigate various approaches to assess contribution of mutations in the emergence and development of drug resistance in Mycobacterium tuberculosis. We present the results of our experiments aimed at identifying resistance mutations to the major anti- TB drugs based on data obtained from patients in Belarus.