检索结果-内蒙古大学图书馆

Proceedings of the 16th IASTED International Conference on Parallel and Distributed Computing and Systems

作者： Panainte, Elena Moscu Beitels, Koen Vassiliadis, Stamatis Computer Engineering Fac. Elec. Eng. Math./Comp. Sci. TU Delft Netherlands

In this paper we consider a set of multimedia applications and investigate the potential performance impact a reconfigurable microcoded processor can provide when added to a general purpose core proeessor. In a design space exploration, considering MPEG2 and JPEG benchmarks, we investigate performance boundaries, memory bottlenecks and the influence the core and reconfigurable processor communication has on performance. Under some realistic scenarios and serial FPGA execution, it is shown that a 53 % cycle reduction is expected when comparing a design having a core processor and a design when the core processor is augmented with a reconfigurable microcoded engine. In addition, we have found that transferring parameters between the core processor and the reconfigurable processor may not severely influence the overall performance. Finally we investigated the memory bandwidth for operations mapped automatically on FPGA. The case study indicates that small latency DCT hardware design performs well when interfac.d with 512 bytes/cycle. Our studies also indicate that about 64 bytes/cycle will support high speed execution for SAD and IDCT.

关键词： Multimedia systems

来源：评论

学校读者我要写书评

暂无评论

A Characterization of Easily Testable Induced Subgraphs

A Characterization of Easily Testable Induced Subgraphs

引用

Proceedings of the Fifteenth Annual ACM-SIAM Symposium on Discrete Algorithms

作者： Alon, Noga Shapira, Asaf Sch. of Math. and Computer Science Raymond/Beverly Sackler Fac. E.S. Tel Aviv University Tel Aviv Israel School of Computer Science Raymond/Beverly Sackler Fac. E.S. Tel Aviv University Tel Aviv Israel

Let H be a fixed graph on h vertices. We say that a graph G is induced H-free if it does not contain any induced copy of H. Let G be a graph on n vertices and suppose that at least en2 edges have to be added to or removed from it in order to make it induced H-free. It was shown in [5] that in this case G contains at least f(Ε,h)nh induced copies of H, where 1/f(Ε, h) is an extremely fast growing function in 1/Ε, that is independent of n. As a consequence, it follows that for every H, testing induced H-freeness with one-sided error has query complexity independent of n. A natural question, raised by the first author in [1], is to decide for which graphs H the function 1/f(Ε,H) can be bounded from above by a polynomial in 1/Ε. An equivalent question is for which graphs H, can one design a one-sided error property tester for testing induced H-freeness, whose query complexity is polynomial in 1/Ε. We settle this question almost completely by showing that, quite surprisingly, for any graph other than the paths of lengths 1, 2 and 3, the cycle of length 4, and their complements, no such property tester exists. We further show that a similar result also applies to the case of directed graphs, thus answering a question raised by the authors in [9]. We finally show that the same results hold even in the case of two-sided error property testers. The proofs combine combinatorial, graph theoretic and probabilistic arguments with results from additive number theory.

关键词： Graph theory

来源：评论

学校读者我要写书评

暂无评论

Centralized matchmaking for minimal agents

Centralized matchmaking for minimal agents

引用

Proceedings of the 16th IASTED International Conference on Parallel and Distributed Computing and Systems

作者： Bertels, K. Panchanathan, N. Vassiliadis, S. Pour Ebrahimi, B. Computer Engineering Lab. Fac. Elec. Eng. Math./Comp. Sci. TU Delft Netherlands

In this paper we propose a simple though efficient mechanism to enable distributed processing nodes to communicate with each other and to dynamically rebalance the workload among them. We discuss a centralized matching mechanism and investigate different matching functions, varying in the amount of information that is taken into account. It is shown that the simplest mechanism, called FirstMatch, results in the fastest matchmaking achieving matching rates of 99%. We also show that this time efficiency does not have a cost in terms of efficient task allocation and resource utilization.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

An experimental study on relationship between pruning algorithms and selection of parameters in fuzzy decision tree generation

An experimental study on relationship between pruning algori...

引用

Proceedings of 2004 International Conference on Machine Learning and Cybernetics

作者： You, Zi-Ying Ji, Hong-Yan Machine Learning Center Fac. of Math. and Computer Science Hebei University Baoding 071002 Hebei China Department of Mathematics Hebei Engineering University Handan 056038 Hebei China

ISBN: (纸本)0780384032

It is important to study the relationship between pruning algorithms and the selection of parameters in fuzzy decision tree generation for controlling the tree size. This paper selects a pruning algorithm and a method of fuzzy decision tree generation to experimentally show the relationship for some existing databases. It aims to give some guidelines for how to select an appropriate parametric value in fuzzy decision tree generation. When a suitable parametric value is selected, the pruning for fuzzy decision tree generation seems to be unnecessary.

关键词： Decision theory

来源：评论

学校读者我要写书评

暂无评论

A personalized products selection assistance based on e-commerce machine learning

A personalized products selection assistance based on e-comm...

引用

Proceedings of 2004 International Conference on Machine Learning and Cybernetics

作者： Yang, Hong-Wei Pan, Zhi-Geng Wang, Xi-Zhao Xu, Bing China State Key Lab. of CAD and CG Zhejiang University Hangzhou 310027 China Machine Learning Center Fac. of Math. and Computer Science Hebei University Baoding 071002 China

ISBN: (纸本)0780384032

The prosperity of electronic commerce has changed the traditional trading behaviors. More and more people are willing to perform Internet shopping. At the same time, consumers experience information overload and look for help in selecting from an overwhelming array of products. In order to overcome such a problem one option is to develop a personalized online assistance to retrieve product information that really matters for the customers. In this paper, we present a method that combines the genetic algorithm and k nearest neighbor technology to reason about the customer's personal preferences from his/her profile and then provide the most appropriate products to meet his/her needs. Our experimental results are showing that our systems have a bright future.

关键词： Electronic commerce

来源：评论

学校读者我要写书评

暂无评论

Case adaptation using estimators of neural network

Case adaptation using estimators of neural network

引用

Proceedings of 2004 International Conference on Machine Learning and Cybernetics

作者： Zhang, Feng Ming-Hu, H.A. Wang, X.I.-Zhao Xiao-Hong, L.I. Machine Learning Center Fac. of Math. and Computer Science Hebei University Baoding 071002 China School of Public Management Hebei Univ. of Econ. and Business shijiazhuang 050054 China

ISBN: (纸本)0780384032

One important issue in Case-Based Reasoning System is the adaptation of cases. In this paper, we proposed a new method for case adaptation, which is based on the estimators of ANN. Using m similar cases to the query case, a Radial Basis Function ANN (RBF network) is trained to provide solutions for the query case. And then an interval solution with certain confidence is obtained by combining the estimators of the ANN model and the proposed solution provided by the RBF network. Experiment shows this approach greatly improves the accuracy of case adaptation by considering the estimators of ANN model.

关键词： Radial basis function networks

来源：评论

学校读者我要写书评

暂无评论

Modelling wind turbines for power system dynamics simulations: An overview

引用

Wind Engineering 2004年第1 SPEC. ISS.期28卷 7-26页

作者： Slootweg, J.G. Kling, W.L. Electrical Power Systems Laboratory Fac. of Elec. Math./Computer Sci. Delft University of Technology P.O. Box 5031 2600 GA Delft Netherlands Cornelis Houtmanstraat 3 8023 EA Zwolle Netherlands

Increasing amounts of wind turbines are connected to electrical power systems in order to mitigate the negative environmental consequences of conventional electricity generation. There are, however, fundamental differences between conventional and wind power generation, particularly regarding the controllability of the prime mover and the generator types used to convert mechanical energy into electricity. These differences relate to the interaction of wind turbines with the power system, including their impact on its dynamic behaviour. In order to investigate the impact of wind power on power system dynamics, adequate models are essential. However, generally accepted models for representing wind turbines in power-system dynamics simulations are presently not generally available. In this paper, various approaches and challenges for the simulation of power system dynamics are discussed and the associated modelling is analysed.

关键词： Electric power systems

来源：评论

学校读者我要写书评

暂无评论

On the comparison of CPLEX-computed job schedules with the self-tuning dynP job scheduler

On the comparison of CPLEX-computed job schedules with the s...

引用

Proceedings - 18th International Parallel and Distributed Processing Symposium, IPDPS 2004 (Abstracts and CD-ROM)

作者： Grothklags, Sven Streit, Achim Fac. Comp. Sci. Elec. Eng./Math. Institute of Computer Science Paderborn University Germany Paderborn Ctr. for Parallel Comp. Paderborn University Germany

ISBN: (纸本)0769521320

In this paper we present a comparison of CPLEX-computed job schedules with the self-tuning dynP scheduler. This scheduler switches the active scheduling policy dynamically during run time, in order to reject changing characteristics of waiting jobs. Each time the self-tuning dynP scheduler checks for a new policy a quasi off-line scheduling is done as the-number of jobs are Lxed. Two questions arise from this fac.: what is the optimal schedule in each self-tuning step? and what is the performance difference between the optimal schedule and the best schedule generated with one of the scheduling policies? For that we modelled the scheduling problem as an integer problem, which is then solved with the well-known CPLEX library. Due to the size of the problem, we apply time-scaling, i. e. the schedule is computed on a larger than one second precise scale. We use the CTC job trace as input for a discrete event simulation and evaluate the performance difference between the CPLEX-computed schedules and the schedules generated by the self-tuning dynP scheduler. The results show, that the performance of the self-tuning dynP scheduler is close to solutions computed by CPLEX. However, the self-tuning dynP scheduler needs much less time for generating the schedules than CPLEX.

关键词： computer systems

来源：评论

学校读者我要写书评

暂无评论

A General Approach to Online Network Optimization Problems

A General Approach to Online Network Optimization Problems

引用

Proceedings of the Fifteenth Annual ACM-SIAM Symposium on Discrete Algorithms

作者： Alon, Noga Awerbuch, Baruch Azar, Yossi Buchbinder, Niv Naor, Joseph Sch. of Math. and Computer Science Raymond/Beverly Sackler Fac. E.S. Tel Aviv University Tel Aviv Israel Johns Hopkins University Baltimore MD 21218 School of Computer Science Raymond/Beverly Sackler Fac. E.S. Tel Aviv University Tel Aviv Israel Computer Science Dept. Technion Haifa 32000 Israel

We study a wide range of online graph and network optimization problems, focusing on problems that arise in the study of connectivity and cuts in graphs. In a general online network design problem, we have a communication network known to the algorithm in advance. What is not known in advance are the bandwidth or cut demands between nodes in the network. Our results include an O(log m log n ) competitive randomized algorithm for the online non-metric fac.lity location and for a generalization of the problem called the multicast problem. In the non-metric fac.lity location m is the number of fac.lities and n is the number of clients. The competitive ratio is nearly tight. We also present an O (log2 n log k) competitive randomized algorithm for the online group Steiner problem in trees and an O(log3n log k) competitive randomized algorithm for the problem in general graphs, where n is the number of vertices in the graph and k is the number of groups. Finally, we design a deterministic O (log3 n log log n) competitive algorithm for the online multi-cut problem. Our algorithms are based on a unified framework for designing online algorithms for problems involving connectivity and cuts. We first present a general O(log m)-deterministic algorithm for generating fractional solution that satisfies the online connectivity or cut demands, where m is the number of edges in the graph. This may be of independent interest for solving fractional online bandwidth allocation problems, and is applicable to the node version as well. The integral solutions are obtained by an online rounding of the fractional solution. This part of the framework is problem dependent, and applies various tools including results on the approximate max-flow min-cut for multicommodity flow, the HST method and its extensions, certain rounding techniques for dependent variables, and Räcke's new hierarchical decomposition of graphs.

关键词： computer networks

来源：评论

学校读者我要写书评

暂无评论

Solving flat-spot problem in back-propagation learning algorithm based on magnified error

Solving flat-spot problem in back-propagation learning algor...

引用

Proceedings of 2004 International Conference on Machine Learning and Cybernetics

作者： Yang, Bo Wang, Ya-Dong Su, Xiao-Hong Wang, Li-Juan Sch. of Comp. Science and Technology Harbin Institute of Technology P.O.BOX 319 Harbin 150001 China Machine Learning Center Fac. of Math. and Computer Science Hebei University Baoding 071002 China

ISBN: (纸本)0780384032

A new learning algorithm based on magnified Error is proposed to speedup the training of back-propagation neural networks, and to improve the performances of neural network. The key to this algorithm lies in varying the error item of output layer, which magnify the backward propagated error signal especially when the weight adjustment of output layer is slow or even suppressed. Therefore, the algorithm is able to get rid of the influence of "flat spot" problem, and solve the slow convergence problem. Consequently the convergence rate can be accelerated, and the training has great capability in meeting the convergence criteria quickly with a simple network structure. Experiments on parity-3 problem and soybean data classification problem show that this method has advantages of faster learning speed and less computational cost than most of the improved algorithms such as sigmoid-prime offset technique (SPO), scaled linear approximation of sigmoid method (SLA) and so on.

关键词： Backpropagation

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：