Digital Commons Network^™

Multigrid Algorithms For Massively Parallel Machines, Satyanarayan Gupta

Computer Science Theses & Dissertations

Numerical solutions of partial differential equations (pde's) are required in many physical problems arising in areas such as computational fluid dynamics, atmospheric sciences, electromagnetics etc. One of the most popular methods of solving pde's is the use of the multigrid algorithm. However, the implementation of the multigrid algorithm on massively parallel machines is not very efficient because of (i) low processor utilization and (ii) high communication overheads. These problems need to be addressed to make better use of massively parallel machines for solving pde's using the multigrid algorithm.

In this dissertation, we present three parallel multigrid algorithms which …

High Performance Issues On Parallel Architectures, Peter J. Looges

Computer Science Theses & Dissertations

In an effort to reduce communication latency in mesh-type architectures, these architectures have been augmented by various types of global and reconfigurable bus structures. The static bus structures provide excellent performance in many areas of computation especially structured numerical computations, but they lack the flexibility required of many large numerical and non-numerical applications. Reconfigurable bus systems have the dynamic adaptability to handle a much wider range of applications. While reconfigurable meshes can often yield constant time results for many problems, the cost of this performance is paid in the number of processors required. While in actuality the majority of these …

Text-Independent Automatic Speaker Identification Using Partitioned Neural Networks, Laszlo Rudasi

Electrical & Computer Engineering Theses & Dissertations

This dissertation introduces a binary partitioned approach to statistical pattern classification which is applied to talker identification using neural networks. In recent years artificial neural networks have been shown to work exceptionally well for small but difficult pattern classification tasks. However, their application to large tasks (i.e., having more than ten to 20 categories) is limited by a dramatic increase in required training time. The time required to train a single network to perform N-way classification is nearly proportional to the exponential of N. In contrast, the binary partitioned approach requires training times on the order of N². …

Parallel Computer Needs At Dartmouth College, David Kotz, Fillia Makedon, Matt Bishop, Scot Drysdale, Don Johnson, Takis Metaxas

Computer Science Technical Reports

To determine the need for a parallel computer on campus, a committee of the Graduate Program in Computer Science surveyed selected Dartmouth College faculty and students in December, 1991, and January, 1992. We hope that the information in this report can be used by many groups on campus, including the Computer Science graduate program and DAGS summer institute, Kiewit's NH Supercomputer Initiative, and by numerous researchers hoping to collaborate with people in other disciplines.

We found significant interest in parallel supercomputing on campus. An on-campus parallel supercomputing facility would not only support numerous courses and research projects, but would provide …

Netsim: A Network Performance Simulator, Lewis Barnett Iii

Department of Math & Statistics Technical Report Series

The performance of computer communication networks is often given only cursory treatment in undergraduate Networking or Computer Communication courses. A simulation package for the investigation of many aspects of Local Area Network (LAN) performance is introduced. Its use as a tool for allowing undergraduate Networking students to investigate the performance of a popular networking system is discussed.

Linear Assignment Problem Revisited: Average (1 + 1/E)-Out Degree Bipartite Graph Has A Perfect Match, Michael Karoński, Wojciech Szpankowski

Department of Computer Science Technical Reports

No abstract provided.

Parallel Iterative Methods, N. P. Christochoides, Elias N. Houstis, S. B. Kim, M. K. Samartzis, John R. Rice

Department of Computer Science Technical Reports

No abstract provided.

Parallel Monte Carlo Trials, Per Brinch Hansen

Electrical Engineering and Computer Science - Technical Reports

The best results of Monte Carlo methods are generally obtained by performing the same computation many times with different random numbers. We develop a generic algorithm for parallel execution of Monte Carlo trials on a multicomputer. The generic algorithm has been adapted for simulated annealing and primality testing by simple substitutions of data types and procedures. The performance of the parallel algorithms was measured on a Computing Surface.

Designing Efficient Maximum-Likelihood Soft-Decision Decoding Algorithms For Linear Block Codes Using Algorithm A*, Yunghsiang S. Han, Carlos R.P. Hartmann

Electrical Engineering and Computer Science - Technical Reports

In this report we present a class of efficient maximum-likelihood soft-decision decoding algorithms for linear block codes. The approach used here is to convert the decoding problem into a search problem through a graph which is a trellis for an equivalent code of the transmitted code. Algorithm A*, which uses a priority-first search strategy, is employed to search through this graph. This search is guided by an evaluation function f defined to take advantage of the information provided by the received vector and the inherent properties of the transmitted code. This function f is used to drastically reduce the search …

Primality Testing, Per Brinch Hansen

Electrical Engineering and Computer Science - Technical Reports

This tutorial describes the Miller-Rabin method for testing the primality of large integers. The method is illustrated by a Pascal algorithm. The performance of the algorithm was measured on a Computing Surface.

Algorithms For Structural Schema Integration, James Geller, Ashish Mehra, Yehoshua Perl, Erich Neuhold, Amit P. Sheth

Kno.e.sis Publications

Current view and schema integration methodologies are driven by semantic considerations, and allow integration of objects only if that is valid from semantic and structural viewpoints. We had introduced a new integration technique called structural integration. It permits integration of objects that have structural similarities, even if they differ semantically. This technique uses the object-oriented Dual Model which separates the representation of structure and semantics.

In this paper we introduce algorithms for structural integration. We apply these algorithms to integrate two views of a large university database schema which had significant structural similarities but differed semantically.

Sequential And Parallel Discrete Event Simulation On Computer Communication Networks, Tao Zhou

Masters Theses

This thesis presents the discrete event simulation technique on the network design, focusing on the methodology, modeling and implementation. The developed technique is suitable for the analysis of any store-and-forward computer communication networks with any level of complexity. The proposed generic simulation model can be used to test the architectures and protocols of existing networks and networks being designed. In implementation of the sequential discrete event simulation, an object-oriented programming strategy is applied to the network simulation using the C++ language. The parallel discrete event simulation is introduced to solve the problem of long time execution of the sequential simulation …

Design Of A Laboratory Computer Interface, Douglas L. Durand

Theses and Dissertations

An interface for the LSI-11 computer was designed and implemented so that the computer supports data acquisition, data reduction, and equipment control. The design includes both hardware and software and addresses both parallel and serial input/output (I/0). The serial interface's hardware is simply a Serial Line Unit card. This card plugs into the LSI-11 bus and provides the signals necessary to interface EIA RS-232 compatible devices. A software utility was developed to allow communication with the serial device and to allow exchange of data files. Routines were written to allow serial I/O through a FORTRAN program. The interface's parallel hardware …

Transformations And Analysis Of Parallel Real Time Programs, Chandima J. Gunasekara

Theses

The problem of schedulability analysis of a set of real time programs form a NP complete problem. The exponential complexity of analysis is a direct result of the complexity in the real time programs, as a combinatorial explosion takes place when trying to determine access patterns of shared resources. Thus, to transform the original programs to a less complex form, while preserving its timing characteristics, is the only viable solution. By using such transformations to reduce the complexity of real time programs, it is possible to schedulability analyze programs at compile time efficiently, without adding an unnecessary overhead to the …

An Implementation And Evaluation Of Distributed Information System Based On The Osca Architecture Guidelines, Babita Masand

Theses

The concepts in this thesis discuss and evaluate the need for systems integration. As a conceptual architecture, the OSCA^TM architecture is investigated. A prototype of the CS Department distributed information system is built following the guidelines and standards of this conceptual architecture. This prototype is implemented in 'C'. RPC (remote procedure call) is used for the communication channel to implement the distributed environment. Conclusions and results that were achieved by implementing this prototype are presented. The main outcome of this thesis is the introduction of the infrastructure to the GenSIF framework.

Graphical Image Persistence And Code Generation For Object Oriented Databases, Subrata Chatterjee

Theses

Attached is the detailed description of the design and implementation of graphical image persistence and code generation for object oriented databases. Graphical image persistent is incorporated into a graphics editor called OODINI. OODINI creates and manipulates graphical schemas for object-oriented databases. This graphical image on secondary storage is then translated into an abstract, generic code for dual model databases. This abstract code, DAL can then be converted into different dual model database languages. We provide an example by generating code for the VODAK Data Modeling language. It is also possible to generate a different abstract language code, OODAL from a …

An Application Of Program Unification To Priority Queue Vectorization, Ling-Yu Chuang, Vernon J. Rego, Aditya P. Mathur

Department of Computer Science Technical Reports

No abstract provided.

Future Research Directions In Problem Solving Environments For Computational Science, Stratis Gallopoulos, Elias N. Houstis, John R. Rice

Department of Computer Science Technical Reports

No abstract provided.

Stabilty Conditions For Some Multiqueue Distributed Systems: Buffered Random Access Systems, Wojciech Szpankowski

Department of Computer Science Technical Reports

No abstract provided.

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 8, Number 5, May 1992, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A six page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Computer Science Education At Wku, Malcolm Stallings

Mahurin Honors College Capstone Experience/Thesis Projects

No abstract provided.

Ua66/3/2 Computer Science Alumni Newsletter, Vol. 7, Wku Computer Science

WKU Administration Documents

Newsletter created by WKU Computer Science for alumni of the program. This issue includes course offerings, computer lab diagram and an alumni survey.

Effects Of Nonsymmetric Release Times On Rate Monotonic Scheduling, R. G. Karl, T. L. Lo, D. C. St. Clair

Computer Science Technical Reports

This paper discusses problems associated with scheduling periodic tasks in a hard, real-time processing or computing environment using a static-priority, preemptive-resume operating system . The scheduling problems associated with a task set containing a single periodic task which has two fixed release periods of unequal length are examined. Some real-world applications may require task release times which are periodic, but whose tasking periods are not symmetric. A scheduling algorithm for task sets with a single nonsymmetric task has been developed for staticpriority, preemptive-resume operating systems. The nonsymmetric scheduling algorithm is based on the rate monotonic scheduling algorithm which assigns higher …

Design Of Backpropagation Neural Network Architectures Using A Decision Tree Classifier, B. M. Van Horn, D. C. St. Clair

Computer Science Technical Reports

The backpropagation neural network algorithm is a popular machine learning methodology. One difficulty with using the algorithm is that the network architecture must be designed before learning can occur. This includes deciding the number of layers, the number of nodes in each layer, and the number of connections. Traditionally this problem is solved by heuristics gained by an expert through experience.

This paper presents an algorithm for using ID3 decision trees to design the network architecture. Previous approaches rely on binary decision trees. The proposed technique utilizes n-ary trees. These trees are easier to understand, are better suited for continuous-valued …

Optimizing Accuracy And Generalization In Numeric Classification Systems, M. D. Walters, D. C. St. Clair

Computer Science Technical Reports

Classifier systems are knowledge-based learning algorithms that take training instances as input and produce a set of rules as output. The classifier systems focused on in this paper represent the knowledge they learn in the form of decision trees, and are built upon Quinlan's [ 1986] inductive algorithm ID3.

The ID3 algorithm suffers from the inability to easily and effectively handle domains with numeric-valued attributes. Numeric attributes are those whose values are taken from a continuous domain or from a domain with a large number of discrete values. A number of approaches have been developed for extending ID3 to handle …

The Identification And Processing Of Don't-Care Attribute Values In Id3 Decision Tree Construction, P. D. Dorr, D. C. St. Clair

Computer Science Technical Reports

ID3 is most successful when used with sets of training and testing data that contain no missing attribute values. Many times, however, real-world domains have attributes with missing values. Sometimes these attribute values may not be needed to classify an instance. Such attribute values are called don't-care attribute values. In other cases, the values are needed but are unavailable. These values are called unknown attribute values. This paper describes the difference between unknown and don't-care attribute values and discusses several ways of identifying don't-care attribute values in ID3. Numerical results are described which validate the practicality of these approaches.

Network Key Management In A Large Distributed Environment, J. J. Stapleton, D. C. St. Clair

Computer Science Technical Reports

The technique of using encryption for protecting information in a network environment involves managing encryption keys within that same network. In large distributed networks the goal of achieving a secure environment requires a secure method of performing network key management. Network security, system security, and application security by means of data encryption rely on encryption keys remaining secret.

Both international and domestic standards organizations such as the International Organization for Standardization (ISO), the American National Standards Institute (ANSI), and the National Institute of Standards and Technology (NIST) address the issues of encryption through various standards. However, these standards discuss methods …

Single Object Detection Using Multiple Sensors With Unknown Noise Distributions, Shaofen Chen

Computer Science Theses & Dissertations

We consider the design of an object classification system that identifies single objects using a system of sensors; each sensor outputs a random vector, according to an unknown (noise) probability distribution, in response to a sensed object. We consider a special class of systems, called the linearly separable systems, where the error-free sensor outputs corresponding to distinct objects can be mapped into disjoint intervals on real line. Given a set of sensor outputs corresponding to known objects, we show that a detection rule α_emp that approaches the correct rule with a high probability can be computed. We show …

Testing A Simple Polygon For Monotonicity Optimally In Parallel, Danny Z. Chen, Sumanta Guha

Department of Computer Science Technical Reports

No abstract provided.

The Development And Application Of Testing Techniques In An Object-Oriented Paradigm, Robert F. Kozal

Masters Theses

This thesis is the study of testing object-oriented programs. Aspects of object-oriented programs for which testing techniques are developed include classes, methods, inheritance, polymorphism, and encapsulation. The application of these techniques are illustrated, and a comparison between traditional testing techniques and object-oriented techniques is also drawn.