检索结果-内蒙古大学图书馆

International Conference on Tools for Artificial Intelligence (ICTAI)

作者： Wenting Zhao Mark H. Liffiton Illinois Wesleyan University Bloomington IL USA

ISBN: (纸本)9781509044603

The problem of enumerating minimal unsatisfiable subsets of a constraint system (MUSes) is a natural candidate for parallelization: as an enumeration problem, it allows for concurrent solving of independent subproblems, and as a typically intractable problem w.r.t. completion (which parallelization cannot transcend), the speed or rate of output (which parallelization can improve) is often the most important performance characteristic. In this work, we explore the parallelization of partial MUS enumeration (aiming to enumerate some MUSes within given resource constraints) via two extensions to a recently-developed sequential algorithm - one employing an existing parallel single-MUS extraction algorithm, the other parallelizing the entire enumeration algorithm-- and we discuss variants and implementation details as well. Results of experiments run with up to 16 cores show that the full parallelization of the entire enumeration algorithm scales well, reaching an average of 92% of perfect scaling with 4 cores and 70% at 16 cores. Evaluating variants and implementation details illuminates how those choices impact performance, including a potentially counterintuitive result that sharing results between threads to avoid duplicate work is not beneficial in the general case.

关键词： Message systems parallel algorithms Feature extraction Runtime Search problems Conferences

来源：评论

学校读者我要写书评

暂无评论

parallel recursive Hamiltonian formulation for constrained multibody system dynamics

Parallel recursive Hamiltonian formulation for constrained m...

引用

2015 ECCOMAS Thematic Conference on Multibody Dynamics, Multibody Dynamics 2015

作者： Chadaj, Krzysztof Malczyk, Pawel Fraczek, Janusz Division of Theory of Machines and Robots Institute of Aeronautics and Applied Mechanics Faculty of Power and Aeronautical Engineering Warsaw University of Technology Nowowiejska 24 Warsaw00-665 Poland

ISBN: (纸本)9788494424403

This paper presents a novel recursive divide and conquer formulation for the simulation of complex constrained multibody system dynamics based on the Hamilton's canonical equations (HDCA). The systems under consideration are subjected to holonomic constraints and may include serial chains, tree chains or closed-loop topologies. Although the Hamilton's canonical equations exhibit many advantageous features compared to their acceleration based counterparts, it appears that there is a lack of dedicated parallel algorithms for multi-rigid body system dynamics based on the Hamiltonian formulation. The developed HDCA formulation leads to a two-stage procedure. In the first phase, the approach utilizes the divide and conquer scheme, i.e. a hierarchic assembly-disassembly process by traversing the multibody system topology in a binary tree manner to evaluate the joint velocities and constraint impulsive loads. The process exhibits linear O{n) (n - number of bodies) and logarithmic O(log2n) numerical cost, in serial and parallel implementations, respectively. The time derivatives of the total momenta are directly evaluated in the second parallelizable step of the algorithm. Sample closed-loop test cases indicate very well constraint satisfaction at the position and velocity level as well as marginal energy drift without any additional form of constraint stabilization techniques involved in the solution process. The results are comparatively set against more standard acceleration based Featherstone's DCA approach to indicate the performance of the HDCA algorithm.

关键词： parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

Infinitely many Zhang functions resulting in various ZNN models for time-varying matrix inversion with link to Drazin inverse

引用

INFORMATION PROCESSING LETTERS 2015年第9期115卷 703-706页

作者： Zhang, Yunong Qiu, Binbin Jin, Long Guo, Dongsheng Yang, Zhi Sun Yat Sen Univ Sch Informat Sci & Technol Guangzhou 510006 Guangdong Peoples R China SYSU CMU Shunde Int Joint Res Inst Foshan 528300 Peoples R China Minist Educ Key Lab Autonomous Syst & Networked Control Guangzhou 510640 Guangdong Peoples R China

In this Letter, by generalizing the notion of Zhang functions (ZFs) from previous work, a novel general-form Zhang function (NGFZF) is proposed, developed and investigated. Specifically, based on the NGFZF, infinitely many ZFs (as error functions) can be readily generated by successively selecting the different values of its parameters. Besides, by employing the NGFZF, a novel general-form Zhang neural net (NGFZNN) is proposed and studied for real-time solution of a time-varying matrix inverse (also termed, Zhang matrix inverse, ZMI). Moreover, a link between ZMI and Drazin inverse is discovered and further generalized to solve for the time-varying Drazin inverse (TVDI). (C) 2015 Elsevier B.V. All rights reserved.

关键词： parallel algorithms Real-time systems Zhang neural net (ZNN) Error functions Zhang matrix inverse (ZMI) Time-varying Drazin inverse (TVDI)

来源：评论

学校读者我要写书评

暂无评论

Microblog data parallel monitoring algorithm on compute unified device Architecture 4th

Microblog data parallel monitoring algorithm on compute unif...

引用

4th International Conference on Computer Engineering and Networks, CENet2014

作者： Cao, Yunpeng Wang, Haifeng College of Information Linyi University LinyiShandong276005 China Institute of LinYi University of Shandong Provincial Key Laboratory of Network Based Intelligent Computing LinyiShandong276005 China

ISBN: (纸本)9783319111032

There is a large-scale information data in microblog systems to be processed in real time. Processing large-scalemicroblog data needs high-performance computing architectures and parallel algorithms. Graphic processing units are adaptable to process data-intensive computing tasks. Using parallel algorithm of GPUs is an effectivemethod to deal with real-time microblog data. Dynamic thread configuration scheme was proposed to decompose the microblog data. The key optimization technique for accessing data on GPUs is solved. Finally, the experimental results show that the proposed parallel algorithm can achieve the average speedup ratio of about 15.3–17.9 and process more than 487 MB per second in practice. © Springer International Publishing Switzerland 2015.

关键词： parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

Stochastic Modeling Simulation to Represent Thermal Physics of Abrasive Machining Based on parallel Computation Technology

Stochastic Modeling Simulation to Represent Thermal Physics ...

引用

World Congress on Engineering (WCE)

作者： D'yakonov, Aleksandr A. South Ural State Univ Dept Technol Engn Chelyabinsk Russia

ISBN: (纸本)9789881404701

Three-dimensional thermo physical model of the abrasive treating processes was formed on the basis of generating integral solving the task in the form of resultant function of the heat source effectiveness with Green function for semi space. Generated model was numerically realized in the form of a software complex for abrasive treating thermo physics simulating stochastic modeling. Software complex structure is divided into three units;each of them is a set of nested loops (nesting level from 2 to 4). Cycles from nesting level 2 allow equivalent transformation to form, containing independent iterations. Upper nesting level cycle allows paralleling under condition of synchronizing input data in the beginning of each iteration. parallel implementation of the software complex uses combination of MPI and OpenMP technologies. Open MP technology is used for decomposition lower level cycles. paralleling efficiency on this level is determined by core cache-memory and by RAM access speed. MPI technology is used for paralleling cycles of upper level and is implemented for calculating with high degree of accuracy. Obtained results demonstrated sufficient calculating time decreasing. For example, 500 reductions under average process input parameters of round grinding with radial feed take approximately 0.5 - 1.2 minutes, while high performance PC processed 50 reductions during 520 minutes.

关键词： Abrasive processing thermo physical model parallel calculations technology parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

Overcomplete dictionary learning with Jacobi atom updates

Overcomplete dictionary learning with Jacobi atom updates

引用

International Conference on Telecommunications and Signal Processing (TSP)

作者： Paul Irofti Bogdan Dumitrescu Department of Automatic Control and Computers University Politehnica of Bucharest Bucharest Romania

ISBN: (纸本)9781509012893

Dictionary learning for sparse representations is traditionally approached with sequential atom updates, in which an optimized atom is used immediately for the optimization of the next atoms. We propose instead a Jacobi version, in which groups of atoms are updated independently, in parallel. Extensive numerical evidence for sparse image representation shows that the parallel algorithms, especially when all atoms are updated simultaneously, give better dictionaries than their sequential counterparts.

关键词： Dictionaries Jacobian matrices Training Optimization Matching pursuit algorithms parallel algorithms Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

A GPU Based Maximum Common Subgraph Algorithm for Drug Discovery Applications

A GPU Based Maximum Common Subgraph Algorithm for Drug Disco...

引用

IEEE International Symposium on parallel and Distributed Processing Workshops and Phd Forum (IPDPSW)

作者： P. B. Jayaraj K. Rahamathulla G. Gopakumar Department of Computer Science and Engineering National Institute of Technology Calicut India

ISBN: (纸本)9781509036837

The maximum common subgraph of two graphs, G 1 and G 2 , is the largest subgraph in G 1 that is isomorphic to a subgraph in G 2 . Finding the maximum common subgraph of two given graphs is known to be a NP-complete problem. An exact solution for the maximum common subgraph problem can be found by an algorithm that transforms the maximum common subgraph problem into a maximal clique enumeration problem. However, as the size of the graph increases, the solution space of the maximal clique enumeration problem increases combinatorially. A serial solution to the computationally intensive problem of complete maximal clique enumeration is tedious. This paper presents a parallel approach using Graphic Processing Unit to compute the maximum common subgraph of the given graphs. The parallel procedure achieves more than tenfold improvement in computational performance. As an application of the proposed parallel maximum common subgraph algorithm, two new tools, LIGANDMATCHER and GRAPHSCREEN are developed. These tools can be used to narrow down the large ligand search space to a small number in the screening phase of drug discovery process.

关键词： Graphics processing units Instruction sets Search problems Kernel parallel algorithms Algorithm design and analysis Drugs

来源：评论

学校读者我要写书评

暂无评论

Load balancing optimization of parallel monte-carlo probabilistic load flow calculation based on openmp

Load balancing optimization of parallel monte-carlo probabil...

引用

International Conference on Power and Energy, CPE 2014

作者： Lu, F. Li, Y.Z. Lu, S.L. Zhang, B.H. Dai, X.K. Wu, J.L. State Grid Jibei Electric Power Company Limited Beijing China State Key Laboratory of Advanced Electromagnetic Engineering and Technology Wuhan China

ISBN: (纸本)9781138027824

The algorithm of probabilistic load flow simulated by serial Monte-Carlo is more timeconsuming due to the more simulation times. It’s difficult to meet the demand for rapid analysis and calculation of large-scale power grid. So it needs to change the serial algorithm into a parallel algorithm in order to shorten the time. But there is the problem of load unbalance with parallel algorithm. This paper first introduces that the probabilistic load flow simulated by Monte-Carlo is more time-consuming. Then based on the OpenMP parallel programming model (Open Multi-Processing, shared memory multithreaded compiler directive), the probabilistic load flow simulated by Monte-Carlo is coarse-grained parallel processing. To deal with the problem of load imbalance in parallel process, we need to optimize the OpenMP parallel programming model through the OpenMP instruction scheduling. With the IEEE162 standard bus system as an example, we simulated the probabilistic load flow based on serial Monte-Carlo for thirty thousand times. The results show that the Load balancing calculated by coarse-grained parallel probabilistic load flow was better, which used OpenMP dynamic scheduling strategy, and we received considerable speedup and parallel efficiency. © 2015 Taylor and Francis Group.

关键词： parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

A parallel algorithm for global states enumeration in concurrent systems 2015

A parallel algorithm for global states enumeration in concur...

引用

20th ACM SIGPLAN Symposium on Principles and Practice of parallel Programming, PPoPP 2015

作者： Chang, Yen-Jung Garg, Vijay K. Department of Electrical and Computer Engineering University of Texas Austin United States

ISBN: (纸本)9781450332057

Verifying the correctness of the executions of a concurrent program is difficult because of its nondeterministic behavior. One of the verification methods is predicate detection, which predicts whether the user specified condition (predicate) could become true in any global states of the program. The method is predictive because it generates inferred execution paths from the observed execution path and then checks the predicate on the global states of inferred paths. One important part of predicate detection is global states enumeration, which generates the global states on inferred paths. Cooper and Marzullo gave the first enumeration algorithm based on a breadth first strategy (BFS). Later, many algorithms have been proposed to improve space and time complexity. None of them, however, takes parallelism into consideration. In this paper, we present the first parallel and online algorithm, named ParaMount, for global state enumeration. Our experimental results show that ParaMount speeds up the existing sequential algorithms by a factor of 6 with 8 threads. We have implemented an online predicate detector using ParaMount. For predicate detection, our detector based on ParaMount is 10 to 50 times faster than RV runtime (a verification tool that uses Cooper and Marzullo's BFS enumeration algorithm). Copyright 2015 ACM.

关键词： parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

High-order accurate difference potentials methods for parabolic problems

引用

APPLIED NUMERICAL MATHEMATICS 2015年 93卷 87-106页

作者： Albright, Jason Epshteyn, Yekaterina Steffen, Kyle R. Univ Utah Dept Math Salt Lake City UT 84112 USA

Highly-accurate numerical methods that can efficiently handle problems with interfaces and/or problems in domains with complex geometry are crucial for the resolution of different temporal and spatial scales in many problems from physics and biology. In this paper we continue the work started in [8], and we use modest one-dimensional parabolic problems as the initial step towards the development of high-order accurate methods based on the Difference Potentials approach. The designed methods are well-suited for variable coefficient parabolic models in heterogeneous media and/or models with non-matching interfaces and with non-matching grids. Numerical experiments are provided to illustrate high-order accuracy and efficiency of the developed schemes. While the method and analysis are simpler in the one-dimensional settings, they illustrate and test several important ideas and capabilities of the developed approach. (C) 2014 Published by Elsevier B.V. on behalf of IMACS.

关键词： Difference potentials Boundary projections Cauchy's type integral Parabolic problems Variable coefficients Heterogeneous media High-order finite difference schemes Difference Potentials Method Immersed Interface Method Interface/composite domain problems Non-matching interface conditions Non-matching grids parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：