Digital Commons Network^™

Multilevel Parallel Communications, Sanjay Khanna

Computer Science Theses & Dissertations

The research reported in this thesis investigates the use of parallelism at multiple levels to realize high-speed networks that offer advantages in throughput, cost, reliability, and flexibility over alternative approaches. This research specifically considers use of parallelism at two levels: the "upper" level and the "lower" level. At the upper level, N protocol processors perform functions included in the transport and network layers. At the lower level, M channels provide data and physical layer functions. The resulting system provides very high bandwidth to an application. A key concept of this research is the use of replicated channels to provide a …

A Framework For Data Sharing In Computer Supported Cooperative Environments, Mohamed Youssef Eltoweissy

Computer Science Theses & Dissertations

Concurrency control is an indispensable part of any information sharing system. Co-operative work introduces new requirements for concurrency control which cannot be met using existing applications and database management systems developed for non-cooperative environments. The emphasis of concurrency control in conventional database management systems is to keep users and their applications from inadvertently corrupting data rather than support a workgroup develop a product together. This "insular" approach is necessary because applications that access the database have been built with the assumptions that they have exclusive access to the data they manipulate and that users of these applications are generally oblivious …

Operational Evaluation Of Responsiveness Properties, Grace Tsai, Matt Insall, Bruce M. Mcmillin

Computer Science Technical Reports

In this paper, a new technique for ensuring run-time satisfaction of properties-specifically responsiveness property, a subset of liveness property, in responsive systems, is presented. Since whether the run-time behavior of a system is satisfied depends on the execution (operational) environment, we develop a translation which takes into account the constraints in the operational environment, and generates histories for each process in the system. Thus, every process can utilize its history to operationally evaluate the system behavior and signal errors if its history is violated. Therefore, this technique provides software safety, handles error-detection, and ensures run-time satisfaction of responsiveness property in …

A General Method For Maximizing The Error-Detecting Ability Of Distributd Algorithms, Martina Schollmeyer, Bruce M. Mcmillin

Computer Science Technical Reports

Error-detecting algorithms can determine when, at run time, a program deviates from its expected behavior due to a hardware, software or communication error. In a fixed interconnect multiprocessor system, the error detecting ability heavily depends on the number of faults, which is bounded, and their spatial distribution. Otherwise multiple fault occurrences can mask each other. This paper provides a general method for computing the overall system failure bound, the maximal fault index, from the system topology and local communication patterns. The result of the computation is used to design a mapping of processes to processor groups such that multiple processor …

Off-Line Cursive Handwriting Recognition Using Style Parameters, Berrin A. Yanikoglu, Peter A. Sandon

Computer Science Technical Reports

We present a system for recognizing off-line cursive English text, guided in part by global characteristics of the handwriting. A new method for finding the letter boundaries, based on minimizing a heuristic cost function, is introduced. The function is evaluated at each point along the baseline of the word to find the best possible segmentation points. The algorithm tries to find all the actual letter boundaries and as few additional ones as possible. After size and slant normalizations, the segments are classified by a one hidden layer feedforward neural network. The word recognition algorithm finds the segmentation points that are …

Ua66/3/2 Computer Science Alumni Newsletter, Vol. 8, Wku Computer Science

WKU Administration Documents

Newsletter created by WKU Computer Science for alumni of the program. This issue includes list of faculty research and scholarly activity and results of alumni survey.

Signal Processing With Neural Networks, Pero Smrzlic

Masters Theses

In this study, we introduce the Adaptive Back Propagation (ABP) learning algorithm, computationally superior to the standard Back Propagation. The ABP is based on the new activation function, with its corresponding adaptive learning parameter. By using the combination of computer simulations and analysis in the domain of activation function, the Method of One Hidden layer was developed for the effective utilization of units in one-hidden layer networks. A parallel version of the ABP was designed and implemented on a nCUBE-2 supercomputer with 128 processors.

The simulation results suggested a strong correlation between frequency of signals and the role of hidden …

Assessment Of The Accuracy Of Three Methods Of Computerized Growth Prediction Of The Soft Tissue Profile, In Untreated Individuals, Pat Diciccio

Loma Linda University Electronic Theses, Dissertations & Projects

This study was performed to determine the accuracy and reliability of the long range growth predictions for profile structures by three computer software systems (QuickCeph™ for Apple Macintosh™ systems. Rocky Mountain Data Systems™ for IBM™ mainframe systems, and Facial Print™ for IBM™ personal computers), on untreated individuals. The total sample consisted of 90 Caucasian children from the Burlington Growth Centre. These were subdivided into groups of 15 children in each of mesiofacial, dolichofacial, and brachyfacial growth types for both males and females, with Angle class I occlusions, normal overbite, and normal overjet. Points measured consisted of 14 skeletal and soft …

Some Experiments With A Basic Linear Algebra Routine On Distributed Memory Parallel Systems, H. Byun, Elias N. Houstis, E. A. Vavalis

Department of Computer Science Technical Reports

No abstract provided.

Feasible Offset And Optimal Offset For Single-Layer Channel Routing, Ronald I. Greenberg, Jau-Der Shih

Computer Science: Faculty Publications and Other Works

The paper provides an efficient method to find all feasible offsets for a given separation in a VLSI channel routing problem in one layer. The prior literature considers this task only for problems with no single-sided nets. When single-sided nets are included, the worst-case solution time increases from Theta(n) to Omega(n^2), where n is the number of nets. But, if the number of columns c is O(n), one can solve the problem in time O(n^{1.5}lg n ), which improves upon a `naive' O(cn) approach. As a corollary of this result, the same time bound suffices to find the optimal offset …

Anns An X Window Based Version Of The Afit Neural Network Simulator, Ching-Seh Wu

Theses and Dissertations

This thesis presents an X Window based neural network simulation environment developed at Air Force Institute of Technology (AFIT) using the techniques of modern software engineering. This artificial neural network simulator is a tool running on Sun SPARCstations and supporting two user modes: end-users and client-programmers. End-users interact with neural network paradigms developed by client-programmers for the purpose of studying and analyzing the execution of a particular Neural Network (NN) paradigm, or class of NN algorithms. Client programmers maintain the system and use this environment for the development of new NN paradigms or algorithms for end-users. The development follows a …

A Decision Criteria To Select An Associative-Memory Organization That Minimizes The Execution Time Of A Mix Of Associative-Search Operations, David W. Banton

Theses and Dissertations

The dissertation develops a decision criteria to select an associative-memory organization that minimizes the execution time of a mix of associative-search operations and a decision criteria to estimate the layout dimensions of each organization for a specified memory size. The dissertation reclassifies Feng's associative-search operations into three hardware-influenced categories: bit-position independent (BPI), record-content independent (RCI); bit-position dependent (BPD), RCI; and BPD, record-content dependent (RCD). It develops a relationship between the categories and three associative-memory organizations: the CAM, the bit-serial word-parallel associative memory (BSWPAM) , and the extreme-search associative memory (ESAM). A version of the CAM, three versions of the BSWPAM, …

Algorithms For Generation Of Path-Methods In Object-Oriented Databases, Ashish Mehta

Dissertations

A path-method is a mechanism in object-oriented databases (OODBs) to retrieve or to update information relevant to one class that is not stored with that class but with some other class. A path-method is a method which traverses from one class through a chain of connections between classes to access information at another class. However, it is a difficult task for a user to write path-methods, because it might require comprehensive knowledge of many classes of the conceptual schema, while a typical user has often incomplete or even inconsistent knowledge of the schema.

This dissertation proposes an approach to the …

Techniques Of Petri Net Reduction, Sreeranga Kalavapalli

Theses

Petri Nets have the capability to analyze large and complex concurrent systems. However, there is one constraint. The number of reachability states of the concurrent systems outweighs the capability of Petri Nets. Previous Petri Net reduction techniques focussed on reducing a subnet to a single transition and hence not powerful enough to reduce a Petri Net. This paper presents six reduction rules and discusses their drawbacks. A new reduction technique called Knitting Technique to delete paths of a Petri Net while retaining all the properties of the original net is presented. Further Structural matrix which facilitates reduction is presented.

Graphical Deep Knowledge Representation In Vodak/Vml Object-Oriented Database, Jue Wang

Theses

Applying Object-oriented concepts to the design of complex graphical interface has received great attention in the database and knowledge representation disciplines. Traditional CAD systems can not support efficient environments for design processes because they store information about all the objects for display purposes but do not store any knowledge for reasoning purposes. They are called "knowledge poor". "Graphical Deep Knowledge" in Artificial Intelligence has been proven successful to represent knowledge about objects for display purposes as well as reasoning purposes. We introduwd the theory of "Graphical Deep Knowledge" into the object-oriented database s stem VML to design a "Knowledge rich" …

Classification Of Patterns In Eeg Recordings : A Comparison Of Back-Propagation Networks Vs. Predictive Autoencoder Networks, Brian Armieri

Theses

Recent research exploring the use of neural networks for electro-encephalogram (EEG) pattern classification has found that a three-layer back-propagation network could be successfully trained to identify high voltage spike-and-wave spindle (HVS) patterns caused by epileptic seizures (Jando et. al., in press). However, there is no reason to predict that back-propagation is the best possible network architecture for EEG classification. A back-propagation neural network and a predictive autoencoder neural network were compared to determine which network was better at correct classifying both HVS and non-HVS patterns.

Both networks were able to classify 88%-89% of all patterns using a limited set of …

Implementation Of An Object-Oriented University Database Using Vodak/Vml Prototype-2/C++, Bheeman Lingan, Madhumathi Tulasiram

Theses

Object-oriented database (OODB) technology has become very popular and successful in recent years. Currently, there are many commercial object-oriented database systems available that are used for developing large and complex real world applications. In addition, there are many research prototypes of object-oriented databases available. In a joint research project of the CIS Dept at NJIT and GMD-IPSI an object-oriented data model called the Dual Model was developed. Using this Dual Model a university environment database schema was designed. This univesity database schema was implemented using the VODAK/VML OODB prototype - 1. The university database was then reimplemented using the C++ …

Throughput Of Existing Multiprocessor File Systems (An Informal Study), David Kotz

Computer Science Technical Reports

Fast file systems are critical for high-performance scientific computing, since many scientific applications have tremendous I/O requirements. Many parallel supercomputers have only recently obtained fully parallel I/O architectures and file systems, which are necessary for scalable I/O performance. Scalability aside, I show here that many systems lack sufficient absolute performance. I do this by surveying the performance reported in the literature, summarized in an informal table.

The State Of Computer Science Facilities Of Schools Across The United States That Are Comparable To Illinois Wesleyan University, Sarah A. Bartz '93

Honors Projects

The purpose of this study is to compare Illinois Wesleyan's academic computer facilities with those at other comparable U.S. colleges and universities. This study also investigates factors affecting these facilities such as the methods other institutions use to deal with and plan for the constantly changing computer world. To obtain this information a survey was mailed to over 250 institutions across the U.S. that were considered comparable to IWU. These institutions include the Associated Colleges of the Midwest, the Oberlin Group, and some of the top colleges and universities in the U.S. as ranked in U.S. News and World Report's …

Diagnosis Of Malignant Melanoma Using A Neural Network, Anurag Chawla, Fikret ErçAl

Computer Science Technical Reports

Malignant melanoma is the deadliest form of all skin cancers. Approximately 32,000 new cases of malignant melanoma were diagnosed in 1991, with approximately 80 percent of patients expected to survive five years [1], Fortunately, if detected early, even malignant melanoma may be treated successfully. Thus, in recent years, there has been a rising interest in the automated detection and diagnosis of skin cancer, particularly malignant melanoma [2]. In this thesis, a novel neural network approach for the automated distinction of melanoma from three benign categories of tumors which exhibit melanoma-like characteristics is presented. The approach is based on devising new …

An Analysis Of Modern Cryptosystems, Thomas Gerald Sauder, Chung You Ho

Computer Science Technical Reports

Since the ancient Egyptian empire, man has searched for ways to protect information from getting into the wrong hands. Julius Caesar used a simple substitution cipher to protect secrets. During World War II, the Allies and the Axis had codes that they used to protect information. Now that we have computers at our disposal, the methods used to protect data in the past are ineffective. More recently, computer scientists and mathematicians have been working diligently to develop cryptosystems which will provide absolute security in a computing environment.

The three major cryptosystems in use today are DES, RSA, and the Knapsack …

Parallel Genetic Algorithms For The Dag Vertex Splitting Problem, Matthias Mayer, Fikret ErçAl

Computer Science Technical Reports

Directed Acyclic Graphs are often used to model circuits and networks. The path length in such Directed Acyclic Graphs represents circuit or network delays. In the vertex splitting problem, the objective is to determine a minimum number of vertices from the graph to split such that the resulting graph has no path of length greater than a given δ. The problem has been proven to be NP-hard.

A Sequential Genetic Algorithm has been developed to solve the DAG Vertex Splitting Problem. Unlike a standard Genetic Algorithm, this approach uses a variable chromosome length to represent the vertices that split the …

The Interpolating Random Spline Cryptosystem And The Chaotic-Map Public-Key Cryptosystem, Fengi Hwu, Chung You Ho

Computer Science Technical Reports

The feasibility of implementing the interpolating cubic spline function as encryption and decryption transformations is presented. The encryption method can be viewed as computing a transposed polynomial. The main characteristic of the spline cryptosystem is that the domain and range of encryption are defined over real numbers, instead of the traditional integer numbers. Moreover, the spline cryptosystem can be implemented in terms of inexpensive multiplications and additions.

Using spline functions, a series of discontiguous spline segments can execute the modular arithmetic of the RSA system. The similarity of the RSA and spline functions within the integer domain is demonstrated. Furthermore, …

Considerations For Rapidly Converging Genetic Algorithms Designed For Application To Problems With Expensive Evaluation Functions, Richard Patrick Rankin, Ralph W. Wilkerson

Computer Science Technical Reports

A genetic algorithm is a technique designed to search large problem spaces using the Darwinian concepts of evolution. Solution representations are treated as living organisms. The procedure attempts to evolve increasingly superior solutions. As in natural genetics, however, there is no guarantee that the optimum organism will be produced.

One of the problems in producing optimal organisms in a genetic algorithm is the difficulty of premature convergence. Premature convergence occurs when the organisms converge in similarity to a pattern which is sub-optimal, but insufficient genetic material is present to continue the search beyond this sub-optimal level, called a local maximum. …

Cartographic Pattern Recognition Using Template Matching, Angela G. Lammers, Ralph W. Wilkerson, Fikret ErçAl

Computer Science Technical Reports

In creating digital maps from paper maps, the paper map must first be scanned to produce a raster image, and then converted into vector format. Vector format allows non-graphical cartographic information to be stored along with the graphical objects. At the United States Geological Survey, the conversion from raster to vector format is performed by a commercial software package. The package also attempts to classify the graphical objects based on shape, line patterns, and other information gained from the raster file. Since the package frequently fails to classify a significant percentage of the elements in the map, manual map analysis …

The Third International Conference On Expert Systems For Numerical Computing, Elias N. Houstis, John R. Rice, Robert Vichnevetsky

Department of Computer Science Technical Reports

No abstract provided.

Broadcast Methods In The Multiswitch Point-To-Point Network, Douglas E. Comer, Victor Norman

Department of Computer Science Technical Reports

No abstract provided.

On The Multisearching Problem For Hypercubes, Mikhail J. Atallah, Andreas Fabri

Department of Computer Science Technical Reports

No abstract provided.

Some Experiments With A Basic Linear Algebra Routine On Distributed Memory Parallel Systems, H. Byun, Elias N. Houstis, E. A. Vavalis

Department of Computer Science Technical Reports

No abstract provided.

The Investigation Of Object-Oriented Methodology As An Enabling Technique For End-User Application Development, Eng Chong Lim

Honors Theses

This research studies the value of object-oriented methods to design information systems that can then be developed and implemented by end-users.