Early operational risk assessment of software using fuzzy expert systems
Zhiwei Xu, Taghi M. Khoshgoftaar, Edward B. Allen · 2003
Operational risk of software is the likelihood of bad events occurring during operations due to software failures. NASA IV&V Facility is an independent institution which conducts independent assessments for various NASA projects. Its responsibilities, among others, include the assessments of operational risks of software systems. In this study, we investigated independent assessments that are conducted very early in the software development life cycle. Existing methods for risk assessment are largely based on checklists and the analysis of risk matrix. Due to the incompleteness of the available software project information during the very early phases of the software life cycle, it is infeasible to use a quantitative method such as risk matrix to assess the operational risk of the software. We developed a fuzzy expert system, called the research prototype early assessment system, to support independent assessments of projects at the very early phases of the software life cycle. We applied the fuzzy expert system to two previous NASA projects, and the results demonstrate that the fuzzy expert system is a very promising technology for early operational risk assessment of software.