Framework for Hybrid Software Evaluation (Classical - Quantum)
Ana Díaz Muñoz
This doctoral dissertation, presented in the form of a compendium of publications, focuses on the design, implementation, and validation of an innovative model for evaluating analyz- ability—understood as a key subcharacteristic of maintainability—in the context of quantum software and, more broadly, hybrid software systems (those integrating classical and quantum components). In a landscape where quantum computing is rapidly advancing and beginning to consolidate in industrial applications, a pressing need arises to adapt and extend traditional software quality models, such as those established in the ISO/IEC 25000 family, to address the specific characteristics of quantum code. The lack of dedicated metrics hinders the rigorous evaluation of critical attributes like analyzability, negatively affecting hybrid systems’ maintainability, evolvability, and reliability in highly complex technical environments. The main objective of this dissertation is the design, implementation, and validation of an evaluation model focused on analyzability, which combines adapted classical metrics with newly defined indicators tailored explicitly for the quantum component. While the model targets hybrid systems, the validation has been exclusively focused on the quantum component, which constitutes the core contribution of this research. The classical part relies on well-established metrics validated and certified by bodies such as ILAC and ENAC and has therefore not been experimentally evaluated in this work. Additionally, an automated technological environment has been developed as a key element of the compendium, facilitating the practical application of the model in real-world projects. This environment enables integrated and real-time measurement of both classical and quantum metrics. However, like the model, its validation has been limited to the novel quantum indicators defined in this thesis. This approach strengthens the proposed methodology and lays the groundwork for future validation of the fully integrated hybrid model. The adopted methodology is grounded in an exhaustive review of the state of the art, the formal and informal definition of metrics, the development of the automated environment, and empirical validation through a family of experiments. The compendium of publications includes, among others, the following articles: Towards a set of metrics for hybrid (quan- tum/classical) systems maintainability, Implementing an environment for hybrid software evaluation, and Environment for the Evaluation of Hybrid (Classical-Quantum) Systems Analyzability. As part of the validation process, a preliminary study has been accepted at the international conference IQSOFT 2025, and an extended version has been submitted to the journal Empirical Software Engineering (EmSE), and is currently under review. The results demonstrate that the proposed model enables precise quantification of quantum software analyzability and that its application in hybrid systems is feasible conceptually and technically. The developed technological environment supports reproducible and systematic evaluation, laying a foundation for future applications in hybrid software engineering. The em- pirical validation conducted thus far has been deliberately limited to the quantum component, given that classical metrics are already widely established and certified. In conclusion, this dissertation represents a significant advance in the evaluation of hybrid software quality. It proposes a validated model that contributes to strengthening maintain- ability and supporting the evolution of complex systems in an ever-changing technological landscape.
•Jan 1, 2025•0 citations