Date of Award

Spring 2002

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Industrial Engineering - (Ph.D.)

Department

Industrial and Manufacturing Engineering

First Advisor

George Hanna Abdou

Second Advisor

Athanassios K. Bladikas

Third Advisor

One-Jang Jeng

Fourth Advisor

Jian Yang

Fifth Advisor

MengChu Zhou

Abstract

Project managers of large software projects, and particularly those associated with Internet Business-to-Business (B2B) or Business-to-Customer (B2C) applications, are under pressure to capture market share by delivering reliable software with cost and timing constraints. An earlier delivery time may help the E-commerce software capture a larger market share. However, early delivery sometimes means lower quality. In addition, most of the time the scale of the software is so large that incremental multiple releases are warranted.

A Multiple-Release methodology has been developed to optimize the most efficient and effective delivery intervals of the various releases of software products, taking into consideration software costs and reliability. The Multiple-Release methodology extends existing software cost and reliability models, meets the needs of large software development firms, and gives a navigation guide to software industrial managers. The main decision factors for the multiple releases include the delivery interval of each release, the market value of the features in the release, and the software costs associated with testing and error penalties. The input of these factors was assessing using Design of Experiments (DOE). The costs included in the research are based on a salary survey of software staff at companies in the New Jersey area and on budgets of software development teams. The Activity Based Cost (ABC) method was used to determine costs on the basis of job functions associated with the development of the software. It is assumed that the error data behavior follows the Non-Homogeneous Poisson Processes (NHPP).

Share

COinS