National Laboratory for Scientific Computing


National Laboratory for Scientific Computing

Time filter
Source Type

Wong T.G.,University of Latvia | Philipp P.,National Laboratory for Scientific Computing
Physical Review A - Atomic, Molecular, and Optical Physics | Year: 2016

Continuous-time quantum walks are natural tools for spatial search, where one searches for a marked vertex in a graph. Sometimes the structure of the graph causes the walker to get trapped, such that the probability of finding the marked vertex is limited. We give an example with two linked cliques, proving that the captive probability can be liberated by increasing the weights of the links. This allows the search to succeed with probability 1 without increasing the energy scaling of the algorithm. Further increasing the weights, however, slows the runtime, so the optimal search requires weights that are neither too weak nor too strong. © 2016 American Physical Society.

Paszynski M.,AGH University of Science and Technology | Azevedo Gomes A.T.,National Laboratory for Scientific Computing
Procedia Computer Science | Year: 2017

In this paper, we present a wrapper around MUMPS solver, called Hierarchical Solver Wrapper (HSW), that is tailored to domain decomposition-based parallel finite element method computations on distributed memory systems. It offers the same interface as parallel MUMPS with matrix entries in coordinate format provided in a distributed fashion among multiple processors. The algorithm implemented by the wrapper utilizes multiple sequential instances of MUMPS solver to compute Schur complements over subdomains. Next, it deallocates sequential MUMPS solvers and LU factors, and it calls the parallel MUMPS solver feeded with the Schur complements, stored in distributed manner. In the backward substitution stage it recomputes the local LU factors before solving the local problems. The wrapper has been tested with three-dimensional isogeometric analysis computations, and we show it reduces both the memory usage and the execution time, in comparison with a single parallel MUMPS call.

Rodrigues P.S.,University of Sao Paulo | Giraldi G.A.,National Laboratory for Scientific Computing
Pattern Analysis and Applications | Year: 2011

Thresholding techniques for image segmentation is one of the most popular approaches in Computational Vision systems. Recently, M. Albuquerque has proposed a thresholding method (Albuquerque et al. in Pattern Recognit Lett 25:1059-1065, 2004) based on the Tsallis entropy, which is a generalization of the traditional Shannon entropy through the introduction of an entropic parameter q. However, the solution may be very dependent on the q value and the development of an automatic approach to compute a suitable value for q remains also an open problem. In this paper, we propose a generalization of the Tsallis theory in order to improve the non-extensive segmentation method. Specifically, we work out over a suitable property of Tsallis theory, named the pseudo-additive property, which states the formalism to compute the whole entropy from two probability distributions given an unique q value. Our idea is to use the original M. Albuquerque's algorithm to compute an initial threshold and then update the q value using the ratio of the areas observed in the image histogram for the background and foreground. The proposed technique is less sensitive to the q value and overcomes the M. Albuquerque and k-means algorithms, as we will demonstrate for both ultrasound breast cancer images and synthetic data. © 2011 Springer-Verlag London Limited.

Lara P.,National Laboratory for Scientific Computing | Borges F.,National Laboratory for Scientific Computing | Portugal R.,National Laboratory for Scientific Computing | Nedjah N.,State University of Rio de Janeiro
Journal of Computer and System Sciences | Year: 2012

The modular exponentiation operation of the current algorithms for asymmetric cryptography is the most expensive part in terms of computational cost. The RSA algorithm, for example, uses the modular exponentiation algorithm in encryption and decryption procedure. Thus, the overall performance of those asymmetric cryptosystems depends heavily on the performance of the specific algorithm used for modular exponentiation. This work proposes new parallel algorithms to perform this arithmetical operation and determines the optimal number of processors that yields the greatest speedup. The optimal number is obtained by balancing the processing load evenly among the processors. Practical implementations are also performed to evaluate the theoretical proposals. © 2011 Elsevier Inc. All Rights Reserved.

Diaz-Caro A.,National University of Quilmes | Yakaryilmaz A.,National Laboratory for Scientific Computing
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2016

We introduce a quantum-like classical computational concept, called affine computation, as a generalization of probabilistic computation. After giving the basics of affine computation, we define affine finite automata (AfA) and compare it with quantum and probabilistic finite automata (QFA and PFA, respectively) with respect to three basic language recognition modes. We show that, in the cases of bounded and unbounded error, AfAs are more powerful than QFAs and PFAs, and, in the case of nondeterministic computation, AfAs are more powerful than PFAs but equivalent to QFAs. © Springer International Publishing Switzerland 2016.

Huang P.G.,Wright State University | Muller L.O.,National Laboratory for Scientific Computing
International Journal for Numerical Methods in Biomedical Engineering | Year: 2015

An extension of a total variation diminishing (TVD) scheme to solve one-dimensional (1D) blood flow for human circulation is proposed. This method is simple as it involves only a few modifications to existing shock-capturing TVD schemes. We have applied the method to a wide range of test cases including a complete simulation of the human vascular network. Excellent solutions have been demonstrated for problems involving varying and discontinuous mechanical properties of blood vessels. For 1D network simulations, the method has been shown to agree well with the reported computational results. Finally, the method has been demonstrated to compare favorably with in vivo experiments set up to study the impact of circle of Willis anomalies on flow patterns in the cerebral arterial system. © 2015 John Wiley & Sons, Ltd.

Todorov M.G.,National Laboratory for Scientific Computing | Fragoso M.D.,National Laboratory for Scientific Computing
Mathematics of Control, Signals, and Systems | Year: 2010

In this paper we introduce the subject of stability radii for continuous-time infinite Markov jump linear systems (MJLS) with respect to unstructured perturbations. By means of the small-gain approach, a lower bound for the complex radius is derived along with a linear matrix inequality (LMI) optimization method which is new in this context. In this regard, we propose an algorithm to solve the optimization problem, based on a bisectional procedure, which is tailored in such a way that avoids the issue of scaling optimization. In addition, an easily computable upper bound for the real and complex stability radii is devised, with the aid of a spectral characterization of the problem. This seems to be a novel approach to the problem of robust stability, even when restricted to the finite case, which in turn allows us to obtain explicit formulas for the stability radii of two-mode scalar MJLS. We also introduce a connection between stability radii and a certain margin of stability with respect to perturbations on the transition rates of the Markov process. The applicability of the main results is illustrated with some numerical examples. © 2010 Springer-Verlag London Limited.

Nakanishi M.,Yamagata University | Yakarylmaz A.,National Laboratory for Scientific Computing
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

In this paper, we show that one-way quantum one-counter automaton with zero-error is more powerful than its probabilistic counterpart on promise problems. Then, we obtain a similar separation result between Las Vegas one-way probabilistic one-counter automaton and one-way deterministic one-counter automaton. Lastly, it was conjectured that one-way probabilistic one blind-counter automata cannot recognize Kleene closure of equality language [A. Yakaryilmaz: Superiority of one-way and realtime quantum machines. RAIRO - Theor. Inf. and Applic. 46(4): 615–641 (2012)]. We show that this conjecture is false. © Springer International Publishing Switzerland 2015.

Goncalves B.,National Laboratory for Scientific Computing | Porto F.,National Laboratory for Scientific Computing
Proceedings of the VLDB Endowment | Year: 2014

In view of the paradigm shift that makes science ever more data-driven, we consider deterministic scientific hypotheses as uncertain data. This vision comprises a probabilistic data- base (p-DB) design methodology for the systematic construc- tion and management of U-relational hypothesis DBs, viz., γ-DB: Managing scientific hypotheses as uncertain data-DBs. It introduces hypothesis management as a promis- ing new class of applications for p-DBs. We illustrate the potential of γ-DB: Managing scientific hypotheses as uncertain data-DB as a tool for deep predictive analytics. © 2014 VLDB Endowment.

Gainutdinova A.,Kazan Federal University | Yakarylmaz A.,National Laboratory for Scientific Computing
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

We continue the systematic investigation of probabilistic and quantum finite automata (PFAs and QFAs) on promise problems by focusing on unary languages. We show that bounded-error QFAs are more powerful than PFAs. But, in contrary to the binary problems, the computational powers of Las-Vegas QFAs and bounded-error PFAs are equivalent to deterministic finite automata (DFAs). Lastly, we present a new family of unary promise problems with two parameters such that when fixing one parameter QFAs can be exponentially more succinct than PFAs and when fixing the other parameter PFAs can be exponentially more succinct than DFAs. © Springer International Publishing Switzerland 2015.

Loading National Laboratory for Scientific Computing collaborators
Loading National Laboratory for Scientific Computing collaborators