Vol 52, No 4 (2018)

A pipelined algorithm for computing the remainder when dividing an arbitrary binary number of a given bit capacity (numerator) by a certain constant value (a constant) is proposed. The algorithm is based on the same types of operations of comparisons and addition–subtraction of partial remainders upon division by this constant. Depending on whether an intermediate result during computation of the remainder is positive or negative, addition with the value of the intermediate result or subtraction from it of the remainder upon division of a given power of two occur. The number of algorithm stages compared with the model is known in advance and depends on the bit capacities of both the dividend and constants. The estimates of the time complexity of the proposed pipelined algorithm are determined by the maximum delay time of operation of the pipeline stage. The estimates of the hardware complexity of the proposed algorithm, as well as the model of the device that implements the algorithm, are determined at the abstract and structural levels.

Nowadays there is an increase in transfer volumes of audio, video and other types of multimedia content in modern computer networks (CN). Multimedia streaming and multicasting applications have limited jitter and end-to-end delay requirements and do not allow full providing of all necessary conditions for efficient data transmission on the Internet. To provide quality of service (QoS) of streaming applications Internet Engineering Task Force (IETF) proposed two QoS architectures: IntServ and DiffServ. However, these architectures have the limited ability of dynamical change in the movement of data flows and do not ensure compliance with all QoS requirements. Complexity in service of CN is often combined with incomplete compatibility of network decisions that involves dependence on the hardware manufacturers. The purpose of the work is the development of the enhanced dynamic load balancing algorithm in CN with QoS. The present work is dedicated to the simulation of software of dynamic load balancing processes. The results of the investigations of the offered algorithm on different topology of CN by 5 QoS-metrics are presented too. The assessment and comparison of the efficiency of the offered load-balancing algorithm with known analog such as Yen’s algorithm including traffic engineering (TE) module are executed.

Mining effective data from wireless sensor network node data is one of the main subjects in studies concerning wireless sensor network data processing. Wireless sensor network data are muli-dimensional and dynamic. Generally, data mining technology cannot satisfy the requirements of wireless sensor network. A large amount of accumulated and redundant wireless sensor network monitoring data reduces the efficiency of data processing. To solve the above problems, this study proposed a data mining algorithm, which integrated rough set algorithm and genetic algorithm to mine redundant data in node network data. The results of the simulated calculation based on MATLAB platform suggested that the identification rate, false accept rate and reject rate of the proposed algorithm were 94.65, 1.753 and 2.331%; compared to network data mining algorithm based on improved genetic algorithm, it has higher efficiency and accuracy in data mining. The algorithm could effectively excavate redundant data in wireless sensor network and optimize the operation environment of wireless sensor network. The application of the rough set and genetic algorithm based data mining algorithm in wireless network has a promising prospect.

In this study we carried out the analysis of the sensitivity of balanced comparators based on small Josephson junction. An expression was obtained for sensitivity in the case of linear growing strobe-pulse of voltage, and it was shown that, sensitivity is determined by the energy ratio-parameter of small size Josephson junction, temperature and sampling rate of comparator. Estimation gives the sensitivity at the level of nV.

The effect of nonideality of the digital-to-analog conversion on properties of the up-anddown method is investigated. Two models for implementation of the up-and-down method, with their own reasons for inaccuracy, are considered and the consequences compared.

The Tangut script was a logographic writing system used for the extinct Tangut language of the Western Xia Dynasty, which spanned 1038 to 1227. The technic of optical character recognition, machine learning, and computer vision will help greatly in the unscrambling of the character in the ancient scripts. But all these technics are based on the character database, which provides learning samples and test standards. In the process of building the Tangut Character Databases using the ancient Tangut scripts as a data source, it is found that the problem of imbalanced class distribution significantly compromises the performance of learning algorithms. A method of synthetic sample generation was proposed in this paper to improve the performance of learning and recognition of Tangut characters. The comparison of recognition accuracy between the learning base in the original data set and the synthetic generated data set was demonstrated, and presented an impressive superiority utilizing the researchers’ method. The organization of Tangut character databases was also introduced in this paper.