Towards Runtime Testing of Dynamically Adaptive Systems based on Behavioral Properties

Abstract

Dynamically Adaptive Systems (DAS) support adaptations to deal with changes in the user requirements and the environments constraints at runtime. A DAS can have high dynamicity of its configurations at runtime, so a major challenge is to perform quality assurance activities. In the literature, approaches that perform runtime testing mostly uses the DAS operational context to generate test plans, then missing other relevant runtime data (e.g., violation of behavioral properties) in the process. Therefore, this master thesis proposes an approach that monitors behavioral properties and use its results to improve the selection and execution of tests at runtime. Thus, the main contribution is the identification of failures in the adaptation mechanism during the DAS execution.