Grover2Q: From UML Models to Executable Qiskit Code: A Model-Driven Engineering Pipeline for Quantum Software
Loading...
Files
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Unuversity Setif 1 Ferhat Abbas . Faculty of Sciences
Abstract
Model-Driven Engineering (MDE) is a software engineering approach that promotes the use of models as first-class artefacts throughout the software development lifecycle, raising the level of abstraction above platform-specific implementation details. The quantum circuit design gets more complex and needs higher levels of abstraction. The development workflows are still largely manual, error-prone and inaccessible to non-experts in quantum mechanics. Although quantum computing platforms such as IBM Qiskit have rapidly matured, there is still a large abstraction gap between high-level system modelling and low-level quantum circuit implementation.
To bridge this abstraction gap, we propose in this Master’s thesis an MDE approach that elevates quantum software development to the model level, separating algorithmic intent from hardware-specific implementation. More precisely, we propose a model transformation from UML Models to Executable Qiskit Code. To this end, we define three metamodels — one for a program's structure, one for its behaviour, and one for the Qiskit target. Then, we propose a chain of two automatic transformations. The first one is an Atlas Transformation Language (ATL) model-to-model step that translates the abstract UML models into a Qiskit model, and the second is an Acceleo (or Python) model-to-text step that turns the obtained model into runnable code. We demonstrate the pipeline on the two-qubit Grover search algorithm. Starting from its class and activity models, the chain produces a working program without a single line of the circuit written by hand, and it recovers the marked state |11⟩ correctly on every run. It is a small but concrete sign that the abstraction and automation of Model-Driven Engineering can carry over to quantum.
