Informatica

Recently, genetic algorithms (GAs) and their hybrids have achieved great success in solving difficult combinatorial optimization problems. In this paper, the issues related to the performance of the genetic search in the context of the grey pattern problem (GPP) are discussed. The main attention is paid to the investigation of the solution recombination, i.e., crossover operators which play an important role by developing robust genetic algorithms. We implemented seven crossover operators within the hybrid genetic algorithm (HGA) framework, and carried out the computational experiments in order to test the influence of the recombination operators to the genetic search process. We examined the one point crossover, the uniform like crossover, the cycle crossover, the swap path crossover, and others. A so-called multiple parent crossover based on a special type of recombination of several solutions was tried, too. The results obtained from the experiments on the GPP test instances demonstrate promising efficiency of the swap path and multiple parent crossovers.

The two major Markov Random Fields (MRF) based algorithms for image segmentation are the Simulated Annealing (SA) and Iterated Conditional Modes (ICM). In practice, compared to the SA, the ICM provides reasonable segmentation and shows robust behavior in most of the cases. However, the ICM strongly depends on the initialization phase.

The paper describes the development and performance of parallel algorithms for the discrete element method (DEM) software. Spatial domain decomposition strategy and message passing inter-processor communication have been implemented in the DEMMAT code for simulation of visco-elastic frictional granular media. The novel algorithm combining link-cells for contact detection, the static domain decomposition for parallelization and MPI data transfer for processors exchanging particles has been developed for distributed memory PC clusters. The parallel software DEMMAT_PAR has been applied to model compacting of spherical particles in the rectangular box. Two benchmark problems with different numbers of particles have been solved in order to measure parallel efficiency of the code. The inter-processor communication has been examined in order to improve domain decomposition topology and to achieve better load balancing. The speed-up equal to 11 has been obtained on 16 processors. The parallel performance study has been performed on the PC cluster VILKAS of Vilnius Gediminas Technical University, Lithuania.

This paper presents an iterative autoregressive system parameter estimation algorithm in the presence of white observation noise. The algorithm is based on the parameter estimation bias correction approach. We use high order Yule–Walker equations, sequentially estimate the noise variance, and exploit these estimated variances for the bias correction. The improved performance of the proposed algorithm in the presence of white noise is demonstrated via Monte Carlo experiments.

In this paper, a digital watermarking algorithm for copyright protection based on the concept of embed digital watermark and modifying frequency coefficients in discrete wavelet transform (DWT) domain is presented. We embed the watermark into the detail wavelet coefficients of the original image with the use of a key. This key is randomly generated and is used to select the exact locations in the wavelet domain in which to embed the watermark. The corresponding watermark detection algorithm is presented. A new metric that measure the objective quality of the image based on the detected watermark bit is introduced, which the original unmarked image is not required for watermark extraction. The performance of the proposed watermarking algorithm is robust to variety of signal distortions, such a JPEG, image cropping, geometric transformations and noises.

This paper presents identity based serial and parallel multisignature schemes using bilinear pairings. Our serial multisignature scheme requires a forced verification at every level to avoid the overlooking of the predecessors' signatures. However, in parallel multisignature scheme the verification of individual signatures is performed by a designated clerk. We show that our schemes are secure against existential forgery under adaptive chosen message attack in the random oracle model.

Determination of stress strain state components of butt welded joint with a mild interlayer at elasto-plastic tension is presented in this paper. Function of normal transverse stresses distribution on the thickness of mild interlayer and equations for determination of the stress intensity maximum value in hard metal at elasto-plastic deforming are proposed. The longitudinal stresses of mild and hard metals at elasto-plastic loading are determined from their integral equilibrium condition to mean value of external load by estimating equality of hard and mild metal longitudinal strains on their contact plane. Algorithm of stress strain state components calculation in separate zones of this model is presented. The strength of butt welded joint with a mild interlayer and longitudinal strain distribution at static and cyclic elasto-plastic loading on its surface calculated analytically is verified by the results of experimental investigations.

Generally, the task in a distributed system must achieve an agreement. It requires a set of processors to agree on a common value even if some components are corrupted. There are significant studies on this agreement problem in a regularized network environment, such as the Fully Connected, BroadCast and MultiCast Networks. Recently, many large complex networks have emerged and displayed a scale-free feature, which influences the system to reach a common value differently. Unfortunately, existing agreement protocols and results cannot cope with the new network environment and the agreement problem thus needs to be revisited. In this paper, we propose a new agreement protocol to adapt to the scale-free network environment and derive its bound of allowable faulty TMs with two rounds of message exchange. We have proved the correctness of this protocol and analyzed its complexity. It is observed that the scale-free network with the proposed agreement protocol can tolerate more faulty TMs than the networks based on previous studies.

New ways to estimate ranges of values of functions from standard and inner interval arithmetic have been proposed. Using the proposed ways ranges of values of mathematical test functions for global optimization and of objective functions for practical global optimization problems have been estimated and compared. Results of the experiments show that it is promising to use proposed balanced interval arithmetic in interval global optimization.

This paper investigates a variety of statistical cache-based language models built upon three corpora: English, Lithuanian, and Lithuanian base forms. The impact of the cache size, type of the decay function, including custom corpus derived functions, and interpolation technique (static vs. dynamic) on the perplexity of a language model is studied. The best results are achieved by models consisting of 3 components: standard 3-gram, decaying cache 1-gram and decaying cache 2-gram that are joined together by means of linear interpolation using the technique of dynamic weight update. Such a model led up to 36% and 43% perplexity improvement with respect to the 3-gram baseline for Lithuanian words and Lithuanian word base forms respectively. The best language model of English led up to a 16% perplexity improvement. This suggests that cache-based modeling is of greater utility for the free word order highly inflected languages.