检索结果-内蒙古大学图书馆

Using Voronoi diagrams to improve classification performances when modeling imbalanced datasets

NEURAL COMPUTING & APPLICATIONS 2015年第5期26卷 1041-1054页

作者： Young, William A., II Nykl, Scott L. Weckman, Gary R. Chelberg, David M. Ohio Univ Coll Business Dept Management Athens OH 45701 USA Ohio Univ Sch Elect Engn & Comp Sci Russ Coll Engn & Technol Athens OH 45701 USA Ohio Univ Dept Ind & Syst Engn Russ Coll Engn & Technol Athens OH 45701 USA

An over-sampling technique called V-synth is proposed and compared to borderline SMOTE (bSMOTE), a common methodology used to balance an imbalanced dataset for classification purposes. V-synth is a machine learning methodology that allows synthetic minority points to be generated based on the properties of a Voronoi diagram. A Voronoi diagram is a collection of geometric regions that encapsulate classifying points in such a way that any point within the region is closest to the encapsulated classifier than any other adjacent classifiers based on their distance from one another. Because of properties inherent to Voronoi diagrams, V-synth identifies exclusive regions of feature space where it is ideal to create synthetic minority samples. To test the generalization and application of V-synth, six databases from various problem domains were selected from the University of California Irvine's Machine Learning Repository. Though not always guaranteed due to the random nature of synthetic over-sampling, significant evidence is presented that supports the hypothesis that V-synth more consistently leads to the creation of more accurate and better-balanced classification models than bSMOTE when the classification complexity of a dataset is high.

关键词： Data mining Knowledge extraction Data engineering numerical algorithms Imbalanced datasets Synthetic over-sampling

来源：评论

学校读者我要写书评

暂无评论

AIAA Policy Statements and Manuscript Review Process Ethical Standards for Publication of Aeronautics and Astronautics Research

引用

Journal of Aircraft 2016年第1期53卷 9-13页

来源：评论

学校读者我要写书评

暂无评论

Safe Positively Invariant Sets for Spacecraft Obstacle Avoidance

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 2015年第4期38卷 720-732页

作者： Weiss, Avishai Petersen, Christopher Baldwin, Morgan Erwin, R. Scott Kolmanovsky, Ilya Univ Michigan Dept Aerosp Engn Ann Arbor MI 48109 USA US Air Force Res Lab Kirtland AFB NM 87117 USA AIAA Reston VA USA

This paper presents an obstacle avoidance method for spacecraft relative motion control. In this approach, a connectivity graph is constructed for a set of relative frame points, which form a virtual net centered around a nominal orbital position. The connectivity between points in the virtual net is determined based on the use of safe positively invariant sets for guaranteed collision free maneuvering. A graph search algorithm is then applied to find a maneuver that avoids specified obstacles and adheres to specified thrust limits. As compared to conventional open-loop trajectory optimization, this approach enables the handling of bounded disturbances, which can represent the effects of perturbing forces and model uncertainty, while rigorously guaranteeing that nonconvex and possibly time-varying obstacle avoidance constraints are satisfied. Details for handling a single stationary obstacle, multiple stationary obstacles, moving obstacles, and bounded disturbances are reported and illustrated with simulation case studies.

关键词： Spacecrafts Model Predictive Control Trajectory Optimization Search Algorithm Fuel Consumption Spacecraft Trajectory Optimization numerical algorithms Spacecraft Maneuvers Lagrange Multipliers Dijkstra's Algorithm

来源：评论

学校读者我要写书评

暂无评论

Relaxation of a precipitate misfit stress state by creep in the matrix

引用

INTERNATIONAL JOURNAL OF PLASTICITY 2015年第1期64卷 164-176页

作者： Fischer, F. D. Svoboda, J. Antretter, T. Kozeschnik, E. Univ Leoben Inst Mech A-8700 Leoben Austria Acad Sci Czech Republic Inst Phys Mat CZ-61662 Brno Czech Republic Vienna Univ Technol Inst Mat Sci & Technol Christian Doppler Lab Early Stages Precipitat A-1040 Vienna Austria

Designed microstructures, such as the ones of high strength and creep resistant modern steels and Al-alloys, are based in many cases on the understanding of the role of precipitates. The kinetics of precipitation can be significantly influenced by the adaption of their volumetric misfit relative to the matrix where plastification and creep of the matrix are typical accommodation processes in addition to vacancy diffusion-controlled mechanisms. A solution for creep coupled with elasticity as such an accommodation mechanism is presented. The set of governing equations including the according assumptions is provided. First, an efficient numerical algorithm is presented to solve the system of two such governing differential equations. Second, necessary steps to obtain an analytical form of the rate equation of the contact pressure are outlined. This rate equation for the contact pressure can be later implemented into models for treatment of precipitation kinetics in complex systems. (C) 2014 Elsevier Ltd. All rights reserved.

关键词： Creep Voids and inclusions Elastic-viscoplastic material numerical algorithms

来源：评论

学校读者我要写书评

暂无评论

On an algorithm for solving parabolic and elliptic equations

引用

COMPUTATIONAL MATHEMATICS AND MATHEMATICAL PHYSICS 2015年第8期55卷 1290-1297页

作者： D'Ascenzo, N. Saveliev, V. I. Chetverushkin, B. N. Russian Acad Sci Keldysh Inst Appl Math Moscow 125047 Russia Deutsch Elektronen Synchrotron DESY D-22206 Hamburg Germany Immanuel Kant Balt Fed Univ Kaliningrad 236041 Russia

The present-day rapid growth of computer power, in particular, parallel computing systems of ultrahigh performance requires a new approach to the creation of models and solution algorithms for major problems. An algorithm for solving parabolic and elliptic equations is proposed. The capabilities of the method are demonstrated by solving astrophysical problems on high-performance computer systems with massive parallelism.

关键词： high-performance computations numerical algorithms parabolic and elliptic equations

来源：评论

学校读者我要写书评

暂无评论

Inexact Newton based Lifted Implicit Integrators for fast Nonlinear MPC

引用

IFAC-PapersOnLine 2015年第23期48卷 32-38页

作者： Quirynen, Rien Gros, Sébastien Diehl, Moritz Department ESAT KU Leuven University Kasteelpark Arenberg 10 Leuven3001 Belgium Signals and Systems Chalmers University of Technology Sweden Department IMTEK University of Freiburg Germany

Nonlinear Model Predictive Control (NMPC) requires the online solution of an Optimal Control Problem (OCP) at every sampling instant. In the context of multiple shooting, a numerical integration is needed to discretize the continuous time dynamics. For stiff, implicitly defined or differential-algebraic systems, implicit schemes are preferred to carry out the integration. The Newton-type optimization method and the implicit integrator then form a nested Newton scheme, solving the optimization and integration problem on two different levels. In recent research, an exact lifting technique was proposed to improve the computational efficiency of the latter framework. Inspired by that work, this paper presents a novel class of lifted implicit integrators, using an inexact Newton method. An additional iterative scheme for computing the sensitivities is proposed, which provides similar properties as the exact lifted integrator at considerably reduced computational costs. Using the example of an industrial robot, computational speedups of up to factor 8 are reported. The proposed methods have been implemented in the open-source ACADO code generation software. © 2015, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.

关键词： Integration Computational efficiency Continuous time systems Model predictive control Newton Raphson method Open source software Open systems Optimal control systems Predictive control systems Continuous time dynamics Differential algebraic systems Newton type optimization Nonlinear model predictive control Nonlinear predictive control numerical algorithms Optimal control problem Optimal controls

来源：评论

学校读者我要写书评

暂无评论

Observation thinning in data assimilation computations

引用

EURO JOURNAL ON COMPUTATIONAL OPTIMIZATION 2015年第1期3卷 31-51页

作者： Gratton, Serge Rincon-Camacho, Monserrat Simon, Ehouarn Toint, Philippe L. Univ Toulouse INP IRIT 2 Rue Camichel F-31000 Toulouse France CERFACS F-31057 Toulouse France Univ Namur Namur Ctr Complex Syst B-5000 Namur Belgium

We propose to use an observation-thinning method for the efficient numerical solution of large-scale incremental four-dimensional (4D-Var) data assimilation problems. This decomposition is based on exploiting an adaptive hierarchy of the observations. Starting with a low-cardinality set and the solution of its corresponding optimization problem, observations are successively added based on a posteriori error estimates. The particular structure of the sequence of associated linear systems allows the use of a variant of the conjugate gradient algorithm which effectively exploits the fact that the number of observations is smaller than the size of the vector state in the 4D-Var model. The new algorithm is tested on a one-dimensional-wave equation and on the Lorenz96 system, the latter one being of special interest because of its similarity with numerical weather prediction systems.

关键词： Data assimilation numerical algorithms Multilevel optimization A posteriori errors

来源：评论

学校读者我要写书评

暂无评论

Iterative Robust Stabilization Algorithm for Periodic Orbits of Hybrid Dynamical Systems: Application to Bipedal Running

引用

IFAC-PapersOnLine 2015年第27期48卷 161-168页

作者： Hamed, Kaveh Akbari Grizzle, Jessy W. Department of Mechanical Engineering San Diego State University San DiegoCA92182 United States Department of Electrical Engineering University of Michigan Ann ArborMI48109 United States

This paper presents a systematic numerical algorithm to design optimal Hoo continuous-time controllers to robustly stabilize periodic orbits for hybrid dynamical systems in the presence of discrete-time uncertainties. A parameterized set of closed-loop hybrid systems is assumed for which there exists a common periodic orbit. The algorithm is created based on an iterative sequence of optimization problems involving Bilinear and Linear Matrix Inequalities (BMIs and LMIs). At each iteration, the optimal %oo problem is translated into a BMI optimization problem which can be easily solved using available software packages. Some sufficient conditions for the convergence of the iterative algorithm are presented. The power of the algorithm is then demonstrated in designing robust stabilizing virtual constraints for running of a highly underactuated bipedal robot with 7 degrees of underactuation in the presence of impact model uncertainties. © 2015, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.

关键词： Iterative methods Chaos theory Continuous time systems Hybrid systems Linear matrix inequalities Optimization Robustness (control systems) Uncertainty analysis Discrete time uncertainty Hybrid dynamical systems Hybrid periodic orbits numerical algorithms Optimization problems Orbital stability Robust stabilization Underactuated

来源：评论

学校读者我要写书评

暂无评论

Two dimensional mathematical model of heat-transmission using MAPLE

引用

IFAC-PapersOnLine 2015年第1期28卷 689-690页

作者： Perjési-Hámori, Ildikó University of Pécs Pécs7624 Hungary

In our poster we show the two dimensional steady-state temperature distribution on rectangular shape surfaces under different boundary conditions. We compare the results of the analytical (Fourier series solution) and that of the numerical (finite-difference method) methods using Maple. We demonstrate the possibility to automatize the numerical algorithm with different boundary conditions. © 2015, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.

关键词： numerical methods Algebra Boundary conditions Finite difference method Fourier series Computer algebra Different boundary condition Heat transmission numerical algorithms Rectangular shapes Series solutions Steady state temperature Two dimensional mathematical models

来源：评论

学校读者我要写书评

暂无评论

Mixing LU and QR factorization algorithms to design high-performance dense linear algebra solvers

引用

JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING 2015年 85卷 32-46页

作者： Faverge, Mathieu Herrmann, Julien Langou, Julien Lowery, Bradley Robert, Yves Dongarra, Jack Univ Bordeaux CNRS Inria Bordeaux INPUMR 5800 Talence France Ecole Normale Super Lyon Lab LIP Lyon France Univ Colorado Denver Denver CO USA Univ Tennessee Knoxville Knoxville TN USA

This paper introduces hybrid LU-QR algorithms for solving dense linear systems of the form Ax = b. Throughout a matrix factorization, these algorithms dynamically alternate LU with local pivoting and QR elimination steps based upon some robustness criterion. LU elimination steps can be very efficiently parallelized, and are twice as cheap in terms of floating-point operations, as QR steps. However, LU steps are not necessarily stable, while QR steps are always stable. The hybrid algorithms execute a QR step when a robustness criterion detects some risk for instability, and they execute an LU step otherwise. The choice between LU and QR steps must have a small computational overhead and must provide a satisfactory level of stability with as few QR steps as possible. In this paper, we introduce several robustness criteria and we establish upper bounds on the growth factor of the norm of the updated matrix incurred by each of these criteria. In addition, we describe the implementation of the hybrid algorithms through an extension of the PaRSEC software to allow for dynamic choices during execution. Finally, we analyze both stability and performance results compared to state-of-the-art linear solvers on parallel distributed multicore platforms. A comprehensive set of experiments shows that hybrid LU-QR algorithms provide a continuous range of trade-offs between stability and performances. (C) 2015 Published by Elsevier Inc.

关键词： numerical algorithms LU factorization QR factorization Stability Performance

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：