TBRIM: decision support for validation/verification of requirements

A decision support system has been developed that provides a structured approach to aid the validation/verification of requirement sets and enhance the quality of the resulting design by reducing risk. Additionally, an automated implementation of this approach has been completed utilizing the Advanced Integrated Requirements Engineering System (AIRES) software. Corroboration for the application of this decision support system has been attained from automated and manual testing performed on the system specification of a large-scale software development and integration project. The decision support approach, called the Test-Based Risk Identification Methodology (TBRIM), has been used to detect four major types of requirements risk potential: ambiguity, conflict, complexity, and technical factors. The TBRIM is based on principles of evidence testing and eliminative logic developed by Bacon and Cohen for making decisions under uncertainty. New techniques for detection of complexity and technical risks have been added to existing methods for identification of risk from ambiguous and conflicting requirements. Comparisons of the automated and manual test results on risk category-by-category basis showed good correlation, and category-independent comparisons showed improvements that were consistent with expectations, Benefits from use of this decision support system include higher accuracy, consistency, and efficiency in the validation/verification of requirements. Knowledge of senior personnel can be captured to provide an expert system for less-experienced personnel