White-Box and Black-Box Test Quality Metrics for Configurable Simulation Models

Abstract

Simulation models are widely employed to model and simulate complex systems from different domains, such as automotive. These systems are becoming highly configurable to support different users' demands. Testing all of them is impracticable, and thus, cost-effective techniques are mandatory. Costs are usually attributed either to the time it takes to test a configurable system or to its monetary value. Nevertheless, for the case of test effectiveness several quality metrics can be found in the literature. This paper aims at proposing both black-box and white-box test quality metrics for configurable simulation models relying on 150% variability modeling approaches.