Informatics and Applications scientific journal
Volume 9, Issue 1, 2015
ANALYTICAL MODELING OF NORMAL PROCESSES IN STOCHASTIC SYSTEMS WITH COMPLEX IRRATIONAL NONLINEARITIES .
Abstract: Stochastic systems (including manifolds) with Wiener and Poisson noises and complex irrational nonlinearities (CIRN) are considered. Equations and algorithms of analytical modeling based on the normal approximation method (NAM) and the statistical linearization method (SLM) are given. Typical integrals and software based on cylindrical functions for computing deterministic and stochastic CIRN are presented. Seven test examples for typical CIRN are given. Applications to Gibbs distributions and distributions with invariant measure are discussed.
Keywords: analytical and statistical modeling; complex irrational nonlinerarity (CIRN); normal approximation method (NAM); statistical linearization method (SLM); test examples
METHOD FOR CALCULATING NUMERICAL CHARACTERISTICS OF TWO DEVICES INTERFERENCE FOR DEVICE-TO-DEVICE COMMUNICATIONS IN A WIRELESS HETEROGENEOUS NETWORK .
Abstract: In wireless networks, one of the key performance metrics is the signal to noise ratio, SINR. As this metric highly depends on the distance between the interfering devices, the problemof SINR estimation is often reduced to the calculation of a triangle’s side length, where the vertices represent the interacting devices. This paper addresses the problem of calculating the numerical characteristics of the signal to interference ratio for a pair of interfering devices determined by the known distributions of distances between the entities in question. The proposed method can be used as a basis for analyzing heterogeneous networks, including the analysis of device-to-device (D2D) communications as one of the interference-limited cases.
Keywords: wireless network; LTE; interference; SINR; D2D
HEURISTIC CERTIFICATES VIA APPROXIMATIONS .
Abstract: This paper suggests a new framework in which the quality of a (not necessarily optimal) heuristic solution is certified by an approximation algorithm. Namely, a result of a heuristic solution is accompanied by a scale obtained from an approximation algorithm. The creation of a scale is efficient while getting a solution from an approximation algorithm is usually concerned with long calculation relatively to heuristics approach. On the other hand, a result obtained by heuristics without scale might be useless. The criteria for choosing an approximation scheme for producing a scale have been investigated. To obtain a scale in practice, not only approximations have been examined by their asymptotic behavior but also relations as a function of an input size of a given problem. For study case only, heuristic and approximation algorithms for the SINGLE KNAPSACK, MAX 3-SAT, and MAXIMUM BOUNDED THREE-DIMENSIONAL MATCHING (MB3DM) NP-hard problems have been examined. The certificates for the heuristic runs have been obtained by using fitting approximations.
Keywords: heuristics; approximation algorithm; optimal solution; approximation preserving reducibility
METHODS AND TOOLS FOR HYPOTHESIS-DRIVEN RESEARCH SUPPORT: A SURVEY .
Abstract: Data intensive research (DIR) is being developed in frame of the new paradigmof research study known as the Fourth paradigm, emphasizing an increasing role of observational, experimental, and computer simulated data practically in all research domains. The principal goal of DIR is an extraction (inference) of knowledge from data. The intention of this work is to make an overview of the existing approaches, methods, and infrastructures of the data analysis in DIR accentuating the role of hypotheses in such process and efficient support of hypothesis formation, evaluation, and selection in course of the natural phenomena modeling and experiments carrying out. An introduction into various concepts, methods, and tools intended for effective organization of hypothesis-driven experiments in DIR is presented.
Keywords: data intensive research; Fourth paradigm; hypotheses; models; theories; hypothetico-deductivemethod; hypothesis testing; hypothesis lattice; Galaxy model; connectome analysis; automated hypothesis generation
FORMAL AXIOMATIC APPROACH TO ASPECT-ORIENTED EXTENSION OF PROGRAMMING TECHNOLOGIES .
Abstract: The procedure of extending modular software systems design technologies by aspect-oriented techniques is considered. The extension is described as enrichment of formal module models by labeling their interfaces by concerns they handle which comprise aspect structure. A novel approach to separation of concerns based on the natural modularizing aspect structure is proposed. Partial modularization of the aspect structure is proposed to generalize this approach. In order to formalize these constructs at the general systems level independently of particular programming paradigms, the category theory is employed. Software engineering technologies are represented as categories with formal models of programs as objects and technological operations as morphisms. The aspect-oriented extension of the technology is axiomatically described as a functor between such categories that has appropriate right and left adjoints. The event-based approach to system modeling is employed as an illustrative case of the aspect-oriented extension.
Keywords: aspect-oriented programming; traceability; category theory; architecture school; separation of concerns
STABLE LINEAR CONDITIONALLY OPTIMAL FILTERS AND EXTRAPOLATORS FOR STOCHASTIC SYSTEMS WITH MULTIPLICATIVE NOISES .
Abstract: The applied theory of analytical synthesis of linear conditionally optimal filters and extrapolators in linear differential stochastic systems with white multiplicative non-Gaussian noises is presented. Efficient criteria of unique asymptotic stability of conditionally optimal filters and extrapolators are formulated in terms of special positive definite integral forms and unique boundedness of controllability and observability matrices. White noises are assumed to be derivatives of additive and multiplicative non-Gausisan arbitrary stochastic processes with independent increments. An illustrative example is given. Some generalizations are discussed.
Keywords: accuracy and unique asymptotic stability of filters; differential stochastic systems; linear conditionally optimal filters and extrapolators; multiplicative white noises; Riccati equation
SELECTION OF OPTIMAL PHYSICAL ACTIVITY CLASSIFICATION MODEL USING MEASUREMENTS OF ACCELEROMETER .
Abstract: The paper solves the problemof selecting optimal stablemodels for classification of physical activity. Each type of physical activity of a particular person is described by a set of features generated froman accelerometer time series. In conditions of feature’s multicollinearity, selection of stable models is hampered by the need to evaluate a large number of parameters of these models. Evaluation of optimal parameter values is also difficult due to the fact that the error function has a large number of local minima in the parameter space. In the paper, the optimal models fromthe class of two-layer artificial neural networks are chosen. The problem of finding the Pareto optimal front of the set of models is solved. The paper presents a stepwise strategy of building optimal stable models. The strategy includes steps of deleting and adding parameters, criteria of pruning and growing the model and criteria of breaking the process of building. The computational experiment compares the models generated by the proposed strategy on three quality criteria — complexity, accuracy, and stability.
Keywords: classification; artificial neural networks; complexity; accuracy; stability; Pareto efficiency; growing and pruning criteria
EVALUATION OF MEASUREMENT ACCURACY AND SIGNIFICANCE FOR LINEAR MODELS .
Abstract: Identification of a linear dependency, when exact solution obtained by standard methods does not meet the objective requirements, determines development of specific approaches for their numerical realization. A method to obtain approximate values of linear models parameters on experimental data, which is based on the use of the linear programmingmethodology and the duality theory, is presented. This method makes it possible to obtain approximate solutions that fulfill all requirements to the model and its parameters and to evaluate accuracy and significance of measurements. It is important for improving the procedure of construction of functional dependencies on the stage of planning experiments if they do not satisfy the authenticity criteria. The results of testing the proposed method for problems connected with research of chemical and socioeconomic systems are given.
Keywords: problems of linear dependencies recovering; measurement accuracy; measurement significance; dual estimates
BAYESIAN RECURRENT MODEL OF RELIABILITY GROWTH:
|Phone: +7 (499) 135-62-60E-mail: email@example.com||RUS|