Cost Prediction for V&V and Certification Processes

Abstract

Cost (time and effort) estimation is fundamental in system and software management. The software related research, together with industries, elaborated and currently use a large number of different cost estimators (CE). While expert judgment is still the most widely used estimation practice in industry, these estimators take a representative calibration set of projects and use a best matching extrapolation curve to predict the costs of further projects. The approach has proven its usefulness in predicting system and software development cost with the breakdown granularity of the main phases of the workflow (specification, implementation, testing etc.). However, there is no such CE for the broad application field of safety critical applications which would be fine granular enough to predict the costs related to Verification and Validation (V&V) and certification. The paper presents an ongoing work for creating such a V&V specific CE. The usefulness of the approach is illustrated by a "what-if" analysis example on the impacts of introducing sophisticated formal methods into the V&V workflow instead of the traditional methodology.