Computer Sciences Commons^™

Comparison Of Scatter Storage Techniques Using An Analysis Of Variance Model, Robert Lee Kitsch

Masters Theses

"An analysis of variance model is developed to determine if a significant difference exists between various scatter storage techniques. The model is a two-factor hierarchical mixed design with each combination of transformation and search method considered as a treatment. The data used in the analysis is obtained from a computer program which provides statistics on the number of probes needed to load the (k+l)st item into a table for the different treatments. An ANOVA table was then computed for various load factors.

A significant difference among the treatments was detected for load factors above .4. Comparison of individual treatments using …

Finite Dimensional Group Rings, Ralph W. Wilkerson

Computer Science Faculty Research & Creative Works

A Ring is Right Finite Dimensional If It Contains No Infinite Direct Sum of Right Ideals. We Prove that If a Group G is Finite, Free Abelian, or Finitely Generated Abelian, then a Ring R is Right Finite Dimensional If and Only If the Group Ring RG is Right Finite Dimensional. a Ring R is a Self-Injective Cogenerator Ring If Rn is Injective and RR is a Cogenerator in the Category of Unital Right /{-Modules; This Means that Each Right Unital A-Module Can Be Embedded in a Direct Product of Copies of R. Let G Be a Finite Group Where …

Monte Carlo Simulation Of Confidence Intervals For Reliability, Kathryn Palisch Berkbigler

Masters Theses

"A computer program which computes fiducial confidence intervals for reliability by Monte Carlo simulation has been written. The program has the capability of computing these intervals for general systems of series, parallel, and mixed series and parallel subsystems which have failure time distributions which are either exponential, Weibull, gamma, normal, or lognormal.

Comparison of the simulation technique with a Bayesian technique for computing confidence intervals in the case of a series of exponentially distributed components shows that the two methods agree quite well when a fiducial prior distribution for reliability is used in the Bayesian technique. A uniform prior distribution …

Computerized Registration For High Schools, Kenneth Lee Fore

Masters Theses

"The guidance personnel in a high school are often burdened with the duty of hand scheduling the student's courses. Using the computer as a sub-optimizing tool in the registration of students, however, can cut scheduling time dramatically and relieve the guidance department of an onerous chore. The technique described in this investigation uses a conflict matrix that schedules the student's request and keeps the class load level within the course sections. A search of the schedule array for each course request may uncover a conflict. If no conflict occurs the course is scheduled and the remaining courses for this student …

Bayesian Confidence Limits For The Reliability Of Mixed Exponential And Distribution-Free Cascade Subsystems, Melvin D. Springer, James K. Byers

Computer Science Faculty Research & Creative Works

The problem treated here is the theoretical one of deriving exact Bayesian confidence intervals for the reliability of a system consisting of some independent cascade subsystems with exponential failure probability density functions (pdf) mixed with other independent cascade subsystems whose failure pdf's are unknown. The Mellin integral transform is used to derive the posterior pdf of the system reliability. The posterior cumulative distribution function (cdf) is then obtained in the usual manner by integrating the pdf, which serves the dual purpose of yielding system reliability confidence limits while at the same time providing a check on the derived pdf. A …

The Assignment Problem And A Suboptimal Solution Technique, William Barry Gregory

Masters Theses

"A technique is presented which finds a fast suboptimal solution to the assignment problem. This same technique is then applied to two large dynamic programming problems for which the optimal solution is not known. These examples illustrate how easily this technique can be applied and that it is better than most optimizing techniques because it is fast, cheap and only minor hand calculations are needed. The effect of different initial solutions and their value are compared and it is found that the initial solution is not as significant in a suboptimal technique as in an optimizing technique. The initial solution …

Modified Method Of Linear Associative Retrieval, James Donald Byrd

Masters Theses

"It is becoming increasingly important for those in decision making positions to have the most current and accurate data possible as related to their particular needs. Although there are several well known retrieval techniques available an intensified effort is being made at many installations to provide a retrieval system that will supply the demands of their own needs. One well known approach to the solution of information retrieval is linear associative retrieval. The solution that has been proposed requires several computer runs on some requests and therefore is very time consuming. The author has implemented a method to eliminate unnecessary …

A Comparison Of Three Numerical Techniques Used For The Solution Of The Two-Point Boundary Value Problem, William Richard Krall

Masters Theses

"Three numerical approximation techniques, called the Initial step, Shooting, and Direct Finite Difference Techniques, respectively, are compared and discussed as they apply to the solution of both linear and non-linear boundary-value problems of the form y"=f(x,y,y') with endpoint conditions y(x_o)=y_o and y(x_n)=y_n. The techniques are compared with respect to speed and accuracy by comparing the solution of each problem considered four times, each time for a different increment value, and comparing the errors at four pivotal points chosen at equally spaced distances over the domain of the function, to give an indication of …

A Development Of Regression Models For Predicting Temperature, Salinity And Sound Velocity In Sea Water, Darwin Wayne Brown

Masters Theses

"There are in existence today some reasonably good numerical procedures for predicting sound velocity, but these procedures require that certain initial data such as temperature and salinity be given before a sound velocity may be computed. Eliminating the need for on-location observations of temperature and salinity, which are used in computing sound velocity, was the primary objective of this study. Stepwise multiple regression was applied in the search for suitable models to predict temperature and salinity, and ultimately sound entirely in terms of controllable variables, such as latitude, longitude, depth, and day-of-year"--Abstract, page ii.

Comparative Analysis Of Polynomial Root Finding Techniques, Charles Ray O'Daniel

Masters Theses

"The purpose of this study was to investigate and recommend various methods instrumental in finding the roots of a polynomial p(x) = 0. Many different methods are present today, and each has its advantages and disadvantages.

Through thorough investigation, the author has ascertained the key methods to be the method of Bisection, the Newton-Raphson method, and the Bairstow method. Special support in the form of algebraic theorems on the locations and kind of roots are extremely helpful. This combination of theorems and methods provides assurance, speed, and the ability to obtain complex roots.

The Bisnewbar method developed by this author …

Stable Predictor-Corrector Methods For First Order Ordinary Differential Equations, Terrell Lester Carlson

Masters Theses

"Because of the wide variety of differential equations, there seems to be no numerical method which will affect the solution best for all problems. Predictor-corrector methods have been developed which utilize more ordinates in the predictor and corrector equations in the search for a better method.

These methods are compared for stability and convergence with the well known methods of Milne, Adams, and Hamming"--Abstract, page ii.

A Comparison Of The Lr And Qr Transformations For Finding The Eigenvalues For Real Nonsymmetric Matrices, Susan Clara Hanson

Masters Theses

"The LR and QR algorithms, two of the best available iterative methods for finding the eigenvalues of a nonsymmetric matrix associated with a system of linear homogeneous equations, are studied. These algorithms are discussed as they apply to the determination of the eigenvalues of real nonsymmetric matrices.

A comparison of the speed and accuracy of these transformations is made. A detailed discussion of the criterion for convergence and the numerical difficulties which may occur in the computation of multiple and complex conjugate eigenvalues are included.

The results of this study indicate that the QR algorithm is the more successful method …

A Simulation Study Of A Proposed Computer Installation, John Edward Hawthorne

Masters Theses

"The operation of a data processing center has been studied with the objective of investigating the effect of various facilities, job loads, and operating policies as measured by job turnaround time. The mechanism for study was digital computer simulation.

This thesis purports to review the technology of simulation and the associated computer techniques, defines the problem and the alternatives available, and analyzes the results derived from the experiments.

Conclusions drawn from this study support the current practice of a functioning data processing center"--Abstract, page ii.

Comparative Analysis Of Numerical Integration Techniques, Edward Lee Sartore

Masters Theses

"When integrating numerically, if the integrand can be expressed exactly as a polynomial of degree n, over a finite interval; then either Simpson's rule, Romberg integration, Legendre-Gauss or Jacobi-Gauss quadrature formulas provide good results. However, if the integrand can not be expressed exactly as an nth degree polynomial, then perhaps it can be expressed as a function f(x) divided by √1-x ², or as a function g(x) times (1-x)^α (l+x)^ß , where α and ß are some real numbers >1, or as a function h(x) times one. If this is the case then the Chebyshev-Gauss, Jacobi-Gauss, and …

A Study Of The Development Of A Heuristic Program That Solves Symbolic Laplace Transformation Problems, Arthur E. Zackrison

Masters Theses

"A medium-speed general purpose digital computer (IBM 1620 Model II) was programmed to solve symbolic Laplace transformation problems. Although the applied mathematician and modern engineer has many uses for the symbolic solution of a problem, the potential of symbolic manipulation by a computer has not yet been sufficiently developed to help them because of many problems. This paper discusses the program and based on the author’s experience with it’s [sic] performance, clarifies the position of several of these difficulties and reaches some conclusions concerning computer solution of such problems"--Abstract, page ii.

Error Analysis Of Rational Approximations Of Functions With Emphasis On Minimax Techniques, Mary Frances Good

Masters Theses

"This study examines the various considerations which are made when a function is approximated by a rational function. None of the four approximations considered here gives both a rapidly calculated approximation and one in which the maximum magnitude of the error function over a given interval is a minimum.

The second algorithm of Remes produces a sequence of rational approximations which converge to the rational approximation that minimizes the maximum magnitude of the error function for a given number of parameters to be calculated if the initial approximation in the sequence is chosen properly. Several rational function approximations are investigated …

A Parameter Perturbation Technique Applied To Multipoint Iteration Functions For The Solution Of Systems Of Nonlinear Equations, Robert N. Delozier

Masters Theses

"The convergence of classical iterative procedures, when applied to a system of nonlinear algebraic or transcendental equations, is highly dependent upon a good initial approximation to the desired roots. Most of the classical iterative schemes have convergence factors between one and two. In this paper iterative schemes of order two and greater are studied in connection with a parameter perturbation process. The parameter perturbation process relaxes the restrictions on the choice of initial values. The procedure divides each problem into a number of subsidiary problems. Each subsidiary system of equations is then solved until a solution is found to the …

Near Optimal Sequencing :N Jobs And M Machines; All Jobs To Be Processed Through The Same Order Of Machines, Larry Glen Clark

Masters Theses

"There is a need, in industry, for an efficient method of determining an optimal sequence for processing a number of jobs through two or more machines. A method which requires a minimum amount of time would be most beneficial. The purpose of this study has been to define such a method. Tests on the method described indicate that although an optimal solution is not always obtained, the solutions obtained are quite good and due to the time factor the method would be beneficial to industry"--Abstract, page ii.