The high-performance learning algorithm for Oja-Karhunen neural network for information compressing is introduced. This learning algorithm allows to process nonlinear and non-stationary data under small sample conditions. The neuro fuzzy architecture of autoassociative memory network is introduced in this paper too. Learning of given construction is based on using .fuzzy basis functions. This hybrid neuro fuzzy system combines advantages of artificial neural networks, fuz y inference systems and evolving systems. The problem of data defining operating lethality clusterization was solved using proposed neuro fuzzy systems.

View full abstract

We develop a new class of positional numeral systems, namely the binomial ones, which form a subclass of generalised positional numeral systems (GPNS). The binomial systems have wide range of applications in the information transmission, processing, and storage due to their error-detection capabilities. In this paper, the binomial numeral systems are well-defined, their prefix and compactness properties are established. Algorithms of generating binomial coding words (non-uniform and uniform) are pre-sented, as well as an enhanced procedure of construction of constant weight Boolean combinations based upon the non-uniform binomial coding words. The correctness of this procedure is also established.

View full abstract

In this paper, we present a model for supply chain design that considers the Cost of Quality as well as the traditional manufacturing and distribution costs. The main contributions are: 1) our model internally computes quality costs for the whole supply chain considering the interdependencies among business entities whereas previous works have assumed exogenously given Cost of Quality functions; and 2) our model can be used at a strategic planning level to design a logistic route that achieves a maximum profit while considering the overall quality level within a supply chain. We provide a solution method based on simulated annealing and discuss several illustrative examples.

View full abstract

The problem of finding a solution to a system of variational inequalities, which can be interpreted as a generalization of a primal-dual variational inequality system and involves non-monotone mappings, is considered. This problem has a great number of applications in Economics, Operations Research, Mathematical Physics and other fields. We suggest this problem to be converted into a variational inequality corresponding to optimality conditions of a non-convex optimization problem. The latter problem is proposed to be solved by a splitting type method.

View full abstract

The methods of rapid information processing aimed to reduce the time of realization of cryptographic RSA transformations are propounded in the paper. These methods are based on application of the principle of ring shift (PRS) in the module number system (MNS). MNS application, from point of view of increasing the speed of realization of cryptographic transformations with the open key, proves to be effective in organizing the process of realization of module integer arithmetic operations.

View full abstract

This paper introduces a Co-Z approach to divisor scalar multiplication in Jacobian of genus 2 hyperelliptic curves over binary fields in projective and weighted coordinates. Approach based on improved addition formulae of weight 2 divisors in projective and weighted divisor representation in most frequent case that suit very well to exponentiation algorithms based on Euclidean addition chains

View full abstract

The method of mathematical modeling for a Bernoulli information source, proposed in [1] first, receive progress in the paper. Bernoulli information sources could be used often for simulation real data sources, which generates information arrays of a different kind inclusive of biomedical ones. When developing methods of data coding for the purpose to increase their efficiency it is pro-posed by us new binomial models for Bernoulli information sources. The considered simulation of a Bernoulli source consists in its separation into two more simple information sources, one of them is combinatory one. As a result it is obtained binomial models of information sources. Such an approach leads to a more simple mathematical description of Bernoulli source models and a more complete reflection of properties of real data sources.

View full abstract

The problem of efficiency of discrete human-machine systems is under consideration in the article The problem of optimum designing of system functioning algorithm is analyzed. A model of selection of option consolidation of functions for a group of operators is offered. Characteristics of error free operation and time of performance of certain operations by human-operators are considered in the paper. An appointment is made according to the compatibility of the operators in the group. For the formulation of the problem we used a model of the algorithm functioning as an events graph.

View full abstract

Set forth here is a way to improve the method of multidimensional scaling by using the second type of the principal components algorithm.

View full abstract

The functional-linguistic approach to the security assurance evaluation is proposed. Methodic and results of the assurance evaluation ontological analysis are given. Ontological modeling application to the security evaluation is grounded.

View full abstract

Perspective directions of expediting the data transfer are identified. A comparative study of methods for identification of traffic in a telecommunications network is conducted. Mathematical models of systems of parametric and structural identification are presented. Advantages and disad-vantages of different methods of identification, as well as opportunities to use these methods for identifying traffic in a telecommunications networks are identified. The ways of further research and development of a method for structural identification of traffic to improve the efficiency of data transmission are offered.

View full abstract

We study the computational complexity of the multi product capacitated lot-sizing problem with pricing. It has been observed that the introduction of prices in the CLSP seems to make the problem computationally easier. The purpose of this paper is to show that even with the introduction of prices the problem is still NP-Hard.

View full abstract