Vol 22, No 2 (2016)

In this paper we describe how prevalent flight routes can be found automatically out of a comprehensive set of arrival flight trajectories. As example k-means clustering of the actual arrival routes is a fundamental step directing to the determination of the airport's prevalent arrival trajectory pencils. For each arrival trajectory pencils then a typical arrival route can be computed, describing the pencil specific mean arrival route (centriod). The polynomial regression model of centroid is useful in this case. As interrelation of centroid and respective trajectory pencils is a sequence of k-means idiom, independent centroid modeling methods for multi dimensional trajectory pencils are under study in present paper. With respect to independent centroid modeling methods, linear and nonlinear dynamical models and Markov models are mentioned. Next, to estimate converging trajectory pencil geometric asymptotes, maximum likelihood orthogonal linear regression models of the trajectory scattered data are in use. The models are obtained by MLESAC algorithm, which is a probabilistic version of RANSAC where the distance of scattered data points from the linear regression model is assumed as distributed according to a mixture of a Gaussian and a uniform distribution. After determining pencil geometric asymptotes sequentially, the trajectory pencils are selected based on trajectory cosine measure to respective asymptotes, and their centroids are defined. As example, trajectory pencil centroids are determined of specific airplane final descending on airport landing-strips of different directions.

Under study is the problem of optimal planning of the partially renewable resources implementation, such as the commercial breeding of fish, animals and so on.

The aim of this article is to study the opportunity of the optimal plan calculation using dynamic programming.

The predetermined scheduling period is divided into a number of stages, such as months or years. At each stage, the key concept of dynamic programming "system status" is formalized as the amount of available resources.

Two models of the objective function are considered. The first model is linear, and the second model is quadratic. On the basis of the dynamic programming method built computational algorithms that do not require busting options step by step. Instead of these classical algorithms of dynamical programming the recurrence formulas for optimal plan calculation are received.

In this paper, the new forecasting method based fuzzy time series, with applying fuzzy implication described, with which help the best result in the decision of forecasting problems received. The suggested method based on relational equations approach, have shown better results of fuzzy time series forecasting. For the proof the known task — forecasting university enrollments of Alabama used. In the paper the comparative analysis of the received results on a background of results received by the above named authors is given. In result, minimal known, that average forecasting error (AFER) in all bases of forecasting in comparison with other methods.

The control technique of system consisting of a set of typical subsystems by computing their ratings is proposed. It is assumed that each subsystem is aimed at increasing their rating.

The control technique includes:

• Selection of a set of indicators subsystems (as well as their distribution by groups).
• Determination of the significance of each indicator to increase the competitiveness of subsystems.
• Setting the significances of each group for a specific period.

The control iteration consists of three steps: setting the initial method of calculating subsystems ratings; implementation the calculated ratings; correction of the manipulated variable (changing the parameters, of subsystems ratings calculations).

Automation control can be implemented in two directions: data collection and calculation of the subsystems ratings and support for the decision to control the global system.

Decision support system are considered and described in terms of automation control. It used database of subsystems and consists of two level. Lower level of decision support system is used for rating computing according with proposed technique. Upper level of decision support system consists of two modules: the module of analytics and the module of forecast. Theoretical estimates of the control system are calculated, among them: estimates of equal and not equal subsystem cases, asymptotical estimate of leader subsystem. The examples of control technique implementation are considered too.

The article presents the results of a study delay QS H₂/M/1 type G/M/1 for a wide range of parameters of traffic. It is known that the distribution in the law Hyperexponential H₂ random variable has a coefficient of variation is greater than 1. It is also known that in the QS G/G/1 waiting time due to the coefficient of variation intervals and service revenues quadratic dependence. Therefore, the waiting time in the QS G/M/1 is also dependent on the coefficient of variation, of arrival intervals. Moreover, it also depends on the higher-order moments. Given the fact that the distribution of H₂ is a three-parameter, the article is an approximation mechanism for arbitrary distribution laws Hyperexponential distribution. This can be done at the level of the first two moments, and at the level of the first three moments. The H₂/M/1 system has that advantage before other systems with input distributions with a heavy tail that for it authors received the exact closed form solution.

An approach of computer aided design (CAD) photo mask workspace layout generation process for very large scale integrated circuit (VLSI) is considered in this paper. The problems of automated CAD approaches lack that occurred during photomask layout design showed. Analysis of photomask construction and workspace designing was considered. Types and photolayers arrangement within photomask workspace is determined. CAD Software architecture and structure development is considered. The main developed software algorithms block-schemes are exposed. Configuration database format and structure for mask layers store of VLSI project arrangement within mask workspace are showed. The developed software integration with Mentor Graphics Calibre CAD system considered. Complex testing of developed CAD software for different types of VLSI layouts (microprocessor and static memory) was done. Analyses of obtained testing results were done. Also timings required for workspace layout generation using developed CAD software were estimated.

The article covers the problem of adopting the synchronous workflow for developing self-timed schemes. The modern EDA tools use the static timing analysis algorithms (STA) as interim step for the most of operations. The major requirement for the STA to work is the absence of feedback interconnections in the analyzed circuit. Suggested approach eliminates the feedbacks in the selftimed pipelines by the modifying the Liberty-model of used elements. The modification affects the functional description of modified elements, so it does not change the arc delays. This allows STA to work, so it becomes possible to use the modern EDA tools to automate design of the self-timed pipelines.

The paper is devoted to the development of a new phase of e-learning (E-Learning 3.0) in which will play an important role: distributed computer systems, cloud computing, mobile personal devices and wireless networks, artificial intelligence and virtual reality tools. Today, there is a fundamental scientific problem of creating a unified interstate intellectual environment for e-education services. This environment should include heterogeneous forms of knowledge representation, interstate standards and forms of learning, international faculty and student teams.

The paper suggests a possible approach to the construction of an all-pervading intellectual environment for e-learning services. The proposed approach is based on the integration of multiple technologies, such as: service-oriented design, engineering of DSPL (Dynamic Software Product Lines), wireless telecommunications, interaction of intelligent e-learning agents. Structural basis of this approach is interconnected pair of intelligent software agents ("avatars"): student’s avatar and teacher’s avatar. These avatars are able to adapt itself: to the current level of student’s knowledge; to the available software, hardware and network equipment; to the current demands of the environment (e.g. labor markets); to the modern educational standards. As a basis of avatars self-adaptation is a variability model, including the three basic features hierarchy: educational content, interface, software and technical support. The mathematical description of this variability model is implemented using the theory of hypergraphs. Application of such variability model does not require recompilation of the source code (to make changes in the avatars properties) and allows to organize a continuous process of e-learning and significantly increase the life cycle of the entire system of distance learning.

Digital health (DH) as a new direction of the healthcare development in the XXI century is based on Biofeedback (BFB). In its turn, fundamental studies of regulation mechanisms of physiological processes in humans and animals provided the basis for the development of biofeedback techniques.

One of the main directions of development of technologies of the digital health is connected with the medical biosensors, sensors of the primary information and wearable devices of microelectronics (WDM), which really enable users of biofeedback system to monitor the status of the organism. The using of biofeedback technologies in the past decade has yielded significant results, when ten thousands humans in the world had the opportunity to take control of their health.

The focus of this work is aimed at the architecture of information system "TsifroMed" which after its implementation will enable the real-time monitor the health of flight crews and air traffic controllers (ie, it is the users of the system, on which depends the safety of flight).

The aim of toe science research program of Moscow aviation institute (MAI) is the creation of information portal "TsifroMed" — digital health, which would provide to users of operational information and provide an access to relevant medical data.

The system "TsifroMed" has a three-tier architecture — an architectural model that assumes the presence of three components: the client (the client layer — the user interface), the application server (the logic layer — software modules and data handler) and database server (data layer — the repository of a large volume data).

The results of this research formed the basis of the project to creation of an information portal "TsifroMed" for on-line communication of participants of the scientific community on the issue of "digital health in aviation and aerospace".

The use case diagram illustrates what actions to perform each member of the project on the portal "TsifroMed". Formed general requirements for information system architecture "Tsifromed".

A conceptual model of the system for further details and documentation has been prepared to communicate with users of the system.

Picked audience of the project and the main modules of the system "TsifroMed", developed the information architecture of the portal. At the stage of implementation developed the database scheme of the system "TsifroMed".