Risk-driven model-based testing of washing machine software: an industrial case study

Abstract

We previously introduced an approach for risk-driven model-based testing. In that approach, test models are represented in the form of Markov chains and test case generation is steered by state transition probabilities. These probabilities are iteratively updated based on estimations of risk of failure. Test cases are generated and executed after each model refinement step to detect new faults in less time. The approach was evaluated in the context of two industrial case studies conducted for testing smart TV and smart phone systems. In this paper, we present results of a third case study, where we apply that approach for model-based testing of washing machine software. We define new types of risks for this particular application, which are estimated based on the number of configuration options and power consumption. As a result, the overall test duration was decreased with risk-driven model-based testing although the number of detected faults did not increase.