Date of Award

5-2012

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Mechanical Engineering

Advisor

Summers, Joshua D.

Committee Member

Mocko , Gregory

Committee Member

Fadel , Georges

Committee Member

Greenstein , Joel

Abstract

Requirements play a critical role within any design process and the activity of identifying and maintaining a system's requirements is essential to. However, design is a complex and iterative process, where requirements are continuously evolving and are volatile. This change, if not managed, may result in undesired uncertainty within the design process leading to monetary losses and time delays, as the changing of requirements has been recognized as a major cause of project failure. In order to mitigate issues that arise due to requirement change propagation, this research presents a computational reasoning tool to help designers and engineers predict change propagation in the requirements domain. The developed tool makes use of requirements syntactical elements to build relationships between requirements. Two heterogeneous industry case studies, spanning four engineering change propagations, are used to both explore the use of requirements in predicting change propagation and generalize an automated prediction tool. Using design structure matrices and graph theoretic based metrics a predictive model is generalized from 491,520 relationship and metric permutation combinations. The developed tool makes use of an RMS scoring algorithm to rank requirements in order of most likely to change due to previous requirement changes. The developed tool is tested against a third industry case study where five engineering changes are predicted. Results indicate the tool can predict sixty percent of change propagation within the top four percent requirements scoring and predict all change propagation within the top thirteen percent scoring.

Share

COinS