Generating Tests for Feature Interaction
Sebastian Benz · 2010
Modern vehicles integrate an increasing number of features. The interaction between these features is often error prone and requires extensive testing in order to ensure correctness and reliability. This thesis presents an approach to test case generation that enables the systematic coverage of feature interactions. A prerequisite to test case generation is a test model that: describes potential error prone paths; and limits the state space to a feasible size. The presented approach uses task models to model features and their interactions. These models are defined in a new task modeling language, TTask. This thesis also introduces new test selection criteria enabling the systematic selection of feature interaction scenarios. The thereby achieved systematic coverage of feature interaction scenarios is demonstrated for real-world systems at BMW Group.