State and prospects of development of the national telecommunication network in the spheres of science and education

The project of creating a modern computer network for Ukrainian educational and scientific institutions is considered, and analysis of the stage of development of this network is performed. The objectives and role of telecommunication networks with INTERNET access at the universities and research institutes of the National Academy of Sciences of Ukraine are shown. The resources of the base center and regional centers are analyzed, the topology and data transmission channels of the network are specified, and the prospects of development of the information infrastructure of the educational and scientific spheres are given. Translated from Kibernetika i Sistemnyi Analiz, No. 5, pp. 3–10, September–October, 1999.     

Nonequilibrium dynamics of a reacting network system

The nonlinear nonequilibrium properties of reacting network systems are studied by computer simulations. It is shown that the fluctuation in the population of each chemical species obeys a log-normal distribution, not the normal Gaussian distribution. The reaction rate shows power-law decay with activation cost (energy), not the Arrhenius-type exponential decay observed in a linear nonequilibrium regime. These two characteristic features will explain the diversity, plasticity, and adaptivity observed in complex biological reaction networks. This work was presented in part at the 13th International Symposium on Artifical Life and Robotics, Oita, Japan, January 31–February 2, 2008     

Synchrony in a self-repair network with a simple lattice

We investigate a synchrony in a self-repairing network of autonomous agents capable of repairing mutually. In this paper, we define two models: a synchronous model and asynchronous one. They differ in the timing when the agents change their state. Computer simulations revealed that the synchronous model has a critical point, while the asynchronous one does not. We also studied a repair scheme in asynchronous model where the repaired agents in turn repair neighbor agents successively in a chain-reactive fashion. Performance of the scheme has been examined by computer simulations. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

A Hopfield neural network approach for power optimization of real-time operating systems

The RTOS (Real-Time Operating System) is a critical component in the SoC (System-on-a-Chip), which is the main body for consuming total system energy. Power optimization based on hardware–software partitioning of a RTOS (RTOS–Power partitioning) can significantly minimize the energy consumption of a SoC. This paper presents a new model for RTOS–Power partitioning, which helps in understanding the essence of the RTOS–Power partitioning techniques. A discrete Hopfield neural network approach for implementing the RTOS–Power partitioning is proposed, where a novel energy function, operating equation and coefficients of the neural network are redefined. Simulations are carried out with comparison to other optimization techniques. Experimental results demonstrate that the proposed method can achieve higher energy savings up to 60% at relatively low costs of less than 4k PLBs while increasing the performance compared to the purely software realized SoC–RTOS.     

A multistage parallel-hierarchic network as a model of a neuronlike computation scheme

A three-dimensional network and its application to the analysis of images are described. This multilevel architecture studies partial correlations between structural components of an image. An algorithm is proposed that formalizes a new approach to the decomposition of images. An image is transformed so that each pixel contains information on the spatial structure of its neighborhood. The most correlated information is first formed, which ensures the resistance of the algorithm to small structural changes. Translated from Kibernetika i Sistemnyi Analiz, No. 2, pp. 114–133, March–April, 2000.     

Using minimal cuts to optimize network reliability for a stochastic computer network subject to assignment budget

In our modern society, information and data are usually transmitted through a computer network. Since the computer network's reliability has a great impact on the quality of data transmission, many organizations devote to evaluating or improving network reliability, especially for network reliability optimization. This study focuses on such a confronted problem that is to find the optimal transmission line assignment to the computer network such that network reliability is maximized subject to the budget constraint. Each transmission line owns several states due to failure, maintenance, etc., and thus the computer network associated with any transmission line assignment is called a stochastic computer network. Network reliability is the probability that the computer network can transmit the specified units of data successfully. Because the discussed problem is NP-hard, an optimization algorithm that integrates the genetic algorithm, minimal cuts and Recursive Sum of Disjoint Products is proposed. Experimental results illustrate the solution procedure and show that the proposed algorithm can be executed in a reasonable time.     

Peer-to-peer sensor network system for a school temperature measurement system

In this paper, we propose the technique of sensor data mining by the peer-to-peer (P2P) network. The mechanism that it is possible to share on the P2P network is considered by receiving information from the sensor by the P2P application. A searching request for a sensor unit and mining the sensor data occurs on the P2P application. We applied the proposed technique to a school environment measurement system. In this system, sensor units are arranged on campus and a user can measure a room’s temperature and humidity. The temperature sensor and the humidity sensor are implemented in a microcomputer board that can connect to the Internet, and we define the microcomputer board as a sensor unit. We construct the P2P sensor network on which a PC accesses the sensor unit and the P2P application on its PC uploads on the P2P network. The P2P network can disclose sensor information after more advanced processing is given by thinking as a P2P application and not a sensor unit, but on the sensor unit and the PC. This work was presented in part at the 12th International Symposium on Artificial Life and Robotics, Oita, Japan, January 25–27, 2007     

Choosing an information coding scheme for a CMAC neural network

The influence of the choice of a coding scheme on properties of a CMAC network is analyzed. Since there exist several arbitrarily chosen parameters in such a network, their optimal values should be found. Computer simulation showed that the type of a coding scheme greatly influences the accuracy of not only identification but also control over nonlinear dynamical objects. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 3, pp. 18–25, May–June 2007.     

Bicriteria network design using a spanning tree-based genetic algorithm

The optimization problems in communication networks have received the attention of many researchers in such related fields as network designer, network analysis, and network administration. The use of computer communication networks has been increasing rapidly in order to share expensive hardware/software resources and provide access to main systems from distant locations. These network problems have many applications in telecommunications, computer networking, and related domains in electric, gas, and sewer networks. In computer networking, LANs (local area networks) are commonly used as the communication infrastructure that meets the demands of users in the local environment. These networks typically consist of several LAN segments connected together via bridges. The use of these transparent bridges requires.loop-free paths between LAN segments. Therefore, only spanning tree topologies can be used as active LAN configurations. Recently, genetic algorithms have greatly advanced in related research fields such as network optimization problems, combinatorial optimization, multiobjective optimization, and so on. Genetic algorithms have also received a great deal of attention because of their ability as optimization techniques for many real-world problems. In this paper, we attempt to solve the LAN topology design problem with bicriteria which minimize the cost and average message delay using genetic algorithms, and propose a method of searching the Pareto solutions. We also employ the Prüfer number in order to represent the chromosomes, because the interconnection between the network service centers must yield spanning tree configurations. Finally, we conduct experiments to certify the quality of the networks designs obtained by using genetic algorithms. This work was presented, in part, at the Third International Symposium on Artificial Life and Robotics, Oita, Japan, January 19–21, 1998     

A face emotion tree structure representation with probabilistic recursive neural network modeling

This paper describes a novel structural approach to recognize the human facial features for emotion recognition. Conventionally, features extracted from facial images are represented by relatively poor representations, such as arrays or sequences, with a static data structure. In this study, we propose to extract facial expression features vectors as Localized Gabor Features (LGF) and then transform these feature vectors into FacE Emotion Tree Structures (FEETS) representation. It is an extension of the Human Face Tree Structures (HFTS) representation presented in (Cho and Wong in Lecture notes in computer science, pp 1245–1254, 2005). This facial representation is able to simulate as human perceiving the real human face and both the entities and relationship could contribute to the facial expression features. Moreover, a new structural connectionist architecture based on a probabilistic approach to adaptive processing of data structures is presented. The so-called probabilistic based recursive neural network (PRNN) model extended from Frasconi et al. (IEEE Trans Neural Netw 9:768–785, 1998) is developed to train and recognize human emotions by generalizing the FEETS representation. For empirical studies, we benchmarked our emotion recognition approach against other well known classifiers. Using the public domain databases, such as Japanese Female Facial Expression (JAFFE) (Lyons et al. in IEEE Trans Pattern Anal Mach Intell 21(12):1357–1362, 1999; Lyons et al. in third IEEE international conference on automatic face and gesture recognition, 1998) database and Cohn–Kanade AU-Coded Facial Expression (CMU) Database (Cohn et al. in 7th European conference on facial expression measurement and meaning, 1997), our proposed system might obtain an accuracy of about 85–95% for subject-dependent and subject-independent conditions. Moreover, by testing images having artifacts, the proposed model significantly supports the robust capability to perform facial emotion recognition.     

Neurolinguistic identification of nonlinear dependences

A method is proposed for the identification of nonlinear dependences on the basis of a composition of a fuzzy knowledge base and a neural network. The structure of a neurofuzzy network that is isomorphic to a system of linguistic statements of the form “if-then” is specified, and analytical models for training are obtained. The efficiency of the proposed method of identification is illustrated by a computer experiment. Translatedfrom Kibernetika i Sistemnyi Analiz, No. 2, pp. 37–44, March–April, 2000     

Performance statistics of a time-sharing computer network

User interaction with a computer is a very complex process and some of the complexity is inherent in the nature of computing. Many experiments in time-sharing and networking of computers have been very extensive and expensive. Nevertheless, a knowledge of the performance and reliability of computer networks is vital for efficient operations. One major problem in the evaluation of any time-sharing network seems to be the lack of quantitative understanding of the relationship for planning future network requirements as well as aiding the process of evaluating existing performance the workload at any computer center is often unique and not typical of a wide variety of commercial environment.This paper presents the following using data collected under actual operating conditions of the LSU time-sharing network.

• 1.a) statistical properties of the input traffic on the network, computer service time demands and selected performance measures.
• 2.b) A model relating to performance measures of the network.

The model developed in this paper consists of two phases: Phase one consists of characterizing the workload performance of time-sharing network by the following model: f(κ) = α + βe^λκ, where α, β and λ are constant coefficients and κ is a performance variable of the network. Phase two consists of developing a regression model of the performance variables of the network. The quality of the model is tested with some real observed data and the difference between predicted values and the observe values are very small. This is supported by the residual analysis of the data.     

Identification of a nonlinear dependence by a fuzzy knowledgebase in the case of a fuzzy training set

This paper generalizes the method of identification of nonlinear dependences by a fuzzy knowledgebase to the case of fuzzy training sets. In such a set, terms are used to estimate inputs. The computer experiments performed showed that the fuzziness of experimental data is no obstacle to identification. The use of fuzzy training sets allows one to apply the proposed method to the identification of "input-output" dependences in medicine, economics, sociology, politics, and other areas in which experimental data are based on expert judgments. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 2, pp. 17–24, March–April 2006. An erratum to this article is available at .     

Stereo-disparity perception for a monochromatic flat slope based on a neural network dynamic model

Stereo-matching is one of the most active research topics in computer vision. In this article, the stereo-correspondence problem for a stereo-image pair on a monochromatic surface is considered. Even if some hints exist, it is not easy to reconstruct the correct 3-D scene from two images because it is an ill-posed problem. We have modified our previous competitive and cooperative neural network model so that we can efficiently perceive a monochromatic surface which is enclosed by two vertical stripes. The modification consists of two factors: (1) combining the parameterized multiple inputs (similarities); (2) extending the cooperative terms of the neural network equation. The effect of the proposed model is examined by experiments with both synthetic and real stereo-image pairs. For the real images, a segmentation method is proposed to deal with the similarity maps. This work was presented, in part, at the Seventh International Symposium on Artificial Life and Robotics, Oita, Japan, January 16–18, 2002     

Algorithm for construction of parametric vectors for solution of classification problems by a feed-forward neural network

A universal heuristic algorithm is considered that constructs the vector of properties of an object that are distributed among heterogeneous groups. This algorithm is based on a feed-forward neural network. An automatic system for objects classification is described. The use of the automatic classification system for a literary authorship attribution problem is considered. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 2, pp. 3–14, March–April 2007.     

Complexity of the problem of verifying the coordination mechanism in a system of software support of network collaboration

The problem of verifying the coordination mechanism in a system of software support of network collaboration is considered. This problem is demonstrated to be similar to the agent verification problem. It is proved that the problem of verifying the coordination mechanism is co-complete. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 4, pp. 15–19, July–August 2008.     

A reinforcement learning method based on an immune network adapted to a semi-Markov decision process

The immune system is attracting attention as a new biological information processing-type paradigm. It is a large-scale system equipped with a complicated biological defense function. It has functions of memory and learning that use interactions such as stimulus and suppression between immune cells. In this article, we propose and construct a reinforcement learning method based on an immune network adapted to a semi-Markov decision process (SMDP). We show that the proposed method is capable of dealing with a problem which is modeled as a SMDP environment through computer simulation. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

An architectural model for intelligent network management

Traditional network management approach involves the management of each vendor‘s equipment and networkd segment in isolation through its own proprietary element management system.It is necessary to set up a new network management architecture that calls for operation consolidation across vendor and technology boundaries.In this paper,an architerctural model for Intelligent Network Management(INM)is presented.The INM system includes a manager system,which controls all subsystems and coordinates different management tasks;an expert system,which is responsible for handling particularly difficult problems,and intelligent agents,which bring the management closer to applications and user requirements by spreading intellignet agents through network segments or domain.In the expert system model proposed,especially an intellignet fault management system is given.The architectural model is to build the INM system to meet the need of managing modern network systems.     

Performance of a closed network with exponential distributed service time and periodically repeating demands of users

A closed network is considered in which one processor provides service to n customers. Service time is exponentially distributed. Once the service is completed, the user calls for the next service in time T. A rejected call returns in time T as well. The performance of the processor and the mean number of call returns are determined. Translated from Kibernetika i Sistemnyi Analiz, No. 3, pp. 176–179, May–June, 2000.     

Stability of an equilibrium state in a multiinfectious-type SIS model on a truncated network

In a homogeneous constant population, the basic SIS model potentially has an epidemic equilibrium state with global asymptotic stability since it can be reduced to the logistic equation. On the basic SIS model with a nonhomogeneous constant population, viewed as a multitype SIS model, the global or local asymptotic stability of an epidemic equilibrium state has also been studied.^1–4 However, this kind of analysis in other models with nonhomogeneous populations has rarely been developed, even though the corresponding models with homogeneous populations are well known. In addition, recent studies of complex networks have revealed that heterogeneity of the link number of vertices drastically changes the epidemic thresholds.^5–9 For these reasons, figuring out the roles of heterogeneity is a major topic in epidemic modeling. Here, we consider a multiinfectious-type SIS model on a network, and show the (local or global) asymptotic stability of an epidemic equilibrium state whenever it exists. This work was presented in part at the 11th International Symposium of Artificial Life and Robotics, Oita, Japan, January 23–25, 2006