Nowadays, software becomes more and more configurable to be adaptable to the ever-growing diversity of requirements and customers’ needs. A key technology to cope with configurable software is software product line engineering, aiming at developing software families consisting of similar but still distinguishable products. Features encapsulate the variability of the functionality of those products. Hence, features denote user-visible configuration options of a software product line (SPL). However, modern software systems not only have to cope with variability, but also with ever-changing conditions such as changing requirements. Dynamic software product lines (DSPL) extend SPLs to model such individually configurable, runtime-adaptive systems being able to reconfigure at runtime.

Especially in the context of mission- and safety-critical software systems, the design of reliable DSPLs requires capabilities for a precise specification of the systems’ functional and reconfiguration behavior. Nevertheless, recent publications related to the specification of DSPLs left the possibility of a precise formulation of complex constraints among configuration parameters and/or respective reconfiguration options out of scope. Besides a precise specification, extensive quality assurance activities for DSPLs with a focus on their reconfiguration capabilities are indispensable to prevent any erroneous behavior of mission- and safety-critical DSPLs at runtime. In practice, software testing is the most established and scalable quality-assurance technique. However, recent work concerning DSPL testing rather provide general testing frameworks expecting a manually created initial test suite as input than a technique to systematically create or even automatically generate DSPL test suites for testing the functional and the reconfiguration behavior of DSPLs.

In this thesis, we present a comprehensive approach for the specification and validation of DSPLs as well as the automated generation of DSPL test suites. To this end, we extend existing DSPL specification approaches by the capabilities to configure features in a flexible staged-configuration process with multiple binding times and to formulate logical as well as temporal dependencies between features by means of complex binding-time constraints. On top of this DSPL specification approach, we further provide automated validation techniques for DSPL specifications to verify essential validity properties of staged configuration processes with complex binding-time constraints during domain engineering.

Furthermore, we present an efficient test-suite generation technique focusing on the reconfiguration behavior of DSPLs. Generating efficient test suites remains computationally expensive even for single products, consisting of repetitive reachability analyses for multiple test goals according to a specific coverage criterion. This situation is even worse when generating test suites for entire software families as apparent in the context of (D)SPLs. Hence, our test-suite generation technique aims at reusing reachability-analysis results to reduce computational efforts at three consecutive levels, the product level, the SPL level and the DSPL level. At product level, we combine existing model-checking-based test-suite generation approaches with recently developed techniques from the area of multi-property checking to increase the reusability of reachability-analysis results among test goals by processing and covering multiple test goals in parallel. At SPL level, this approach is extended by the capability to generate test suites for an entire software family by reusing reachability-analysis results among products. Finally, we extend the approach at DSPL level by considering the reconfiguration behavior of DSPLs beside their functional behavior.

The concepts and techniques contributed in this thesis are primarily motivated and illustrated by means of a real-world high-configurable runtime-adaptive system from the automation domain. Based on prototypical implementations, we provide an evaluation of the presented techniques using various real-world case studies.