VTechWorks staff will be away for the winter holidays starting Tuesday, December 24, 2024, through Wednesday, January 1, 2025, and will not be replying to requests during this time. Thank you for your patience, and happy holidays!

Browsing by Author "Hirschman, Edward"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Comprehensive forecasting of software integrity in C3I systems
    Hirschman, Edward (Virginia Tech, 1992-10-15)
    The purpose of this project is to forecast the incidence of failures to be encountered by a software package for C3I systems over and throughout its life cycle. It will be assumed that a data base of software previously developed for C3I systems will be used to forecast the software integrity of a software package under initial development. "Software integrity" is defined as a projection of the stream of failures that will be experienced by the new software. The failure history of the mature C3I systems software will be statistically quantified parametrically and by experimental design techniques (ANOVA) to gather information which will be used to forecast what C3I software with similar characteristics--length, language, debugging effort, etc.--will experience. Then, as the new C3I system software matures, statistical techniques for software systems engineering will be addressed for testing appropriateness of the initial projections; and eventually the new software will be parametrically modeled on its own merits to forecast the failures to be encountered over the remainder of its life cycle. Lastly, the data base history of software for mature C3I systems software will be updated and amended as needed to facilitate reliable forecasting of software integrity for a new round of C3I systems software. The attention to C3I implied by the title of the project will reflect itself in the classes of software considered and development conditions, schedules and complexities of the software.
  • Thumbnail Image
    Optimal class scheduling subject to professors' preferences
    Hirschman, Edward (Virginia Tech, 1989-03-05)
    This new form of multiattribute utility optimization is based on ordinal as opposed to cardinal utility and is defined from a corresponding integer programming model in operations research which (1) is solved for ordinal cost factors and (2) serves as the problem's theoretical starting point. It is suggested herein that one start with a mathematical formulation that if solved in an acceptable or — preferably — best manner would yield a satisfactory or possibly best solution to the problem. Then, that mathematical formulation and its solution technique defines the multiattribute utility problem and its solution at issue. This is the reverse of what is usually done; and as will be shown, doing this can be quite fruitful. The illustrative example concerns a mathematical 1 formulation based on operation research's assignment problem. As will be argued, the cost factors must be ordinal, which essentially corresponds to using ordinal utility; hence the technique will be framed in the realm of ordinal utility. The technique for solving the illustrative example's mathematical formulation is to achieve a premium mix of operations research solution properties. From this perspective, some sticky issues in multiattribute utility theory when the attributes involve the preferences of distinct persons are not included in the philosophical base for the multiattribute utility problem and its solution thusly defined.