Thesis defense of Urtzi Markiegi Gonzalez

Thesis defense of Urtzi Markiegi Gonzalez

Title of the thesis: "Test optimization for Highly-Configurable Cyber-Physical Systems". Obtained the SOBRESALIENTE CUM LAUDE qualification.

---

• Title of the thesis: "Test optimization for Highly-Configurable Cyber-Physical Systems".
• Court:
  • President: Óscar Díaz García (UPV/EHU)
  • Vocal: Shaukat Ali (Simula Research Laboratory)
  • Vocal: David Benavides Cuevas (Universidad de Sevilla)
  • Vocal: Salvador Trujillo González (Ikerlan)
  • Secretary: Aitor Arrieta Marcos (Mondragon Unibertsitatea)

Abstract

Over the past decade, Cyber-Physical Systems (CPSs) have gained prominence as core-enabling technologies in the development of multiple domains, thanks to its ability to integrate digital capabilities with physical processes. Furthermore, the demand for configurability of CPSs has been increasing rapidly to respond to changing business requirements.

When engineers approach the development of Highly-Configurable Cyber-Physical Systems (HCCPSs), product line engineering techniques are often adopted, taking advantage of variability management strategies that allow handling a large number of configurations. In addition, to address intrinsic challenges of testing CPSs, engineers are relying on simulation-based techniques, thus avoiding the need for building real prototypes and enabling testing at early stages. However, HCCPS testing is still a time-consuming challenge primarily due to the intensive resource consumption when simulating physical processes. Consequently, optimization of testing HCCPSs is paramount.

Several approaches have tackled the test optimization challenge, most of them focused on reducing the number of products to be tested, by selecting a representative subset. Other approaches have proposed optimization in terms of test case selection and prioritization. However, little attention has been paid to optimization of both, products and test cases, in a combined manner. This thesis aims at advancing the current practice of optimizing HCCPS testing by proposing a method to increase the fault detection rate in time-constrained scenarios. To this end, we propose a dynamic test prioritization approach combining both, products and test cases. The approach sets a test plan that executes small groups of test cases with products in iterative executions. After every iteration, the test plan is dynamically re-ordered, leveraging information of test cases being executed in specific products. The approach has been evaluated and validated for the specific context of HCCPSs, however, it might eventually pave the way for its use in other type of product lines.