Quantum Computing Offers Potential Gains for Software Engineering Tasks.

Quantum-based software engineering (QBSE) presents a novel approach to enhance computationally demanding software tasks including testing, analysis, and defect prediction. Research indicates potential benefits from applying quantum optimisation, search, and learning techniques, though current efforts lack cohesion and require a focused research agenda.

The escalating demands placed on software systems necessitate continual improvements in development and maintenance processes. Researchers are now investigating whether the principles of quantum computation can offer advantages in tackling computationally challenging software engineering tasks. Jianjun Zhao from Kyushu University, and colleagues, explore this potential in a new analysis of ‘Quantum-Based Software Engineering’ (QBSE). Their work distinguishes QBSE from simply using quantum computers for software tasks (Quantum Software Engineering, or QSE), instead focusing on adapting quantum algorithms – particularly those relating to optimisation, search and machine learning – to enhance established software engineering methodologies like test case selection, static analysis and defect prediction. The paper consolidates existing, disparate research and proposes a focused research agenda to further develop this emerging field.

Quantum Software Engineering: Beyond Hardware Control

Quantum-based software engineering (QBSE) investigates the application of quantum computing principles to improve traditional software development processes, distinct from utilising software to simply control quantum hardware. Current research centres on three primary areas: optimisation of software testing, application of quantum machine learning (QML), and enhancement of software analysis techniques.

Researchers employ quantum algorithms, notably Quantum Annealing and the Quantum Approximate Optimisation Algorithm (QAOA), to minimise test suite sizes and select the most effective test cases. This aims to reduce computational cost and improve the efficiency of regression testing – a critical component of software maintenance where existing functionality is re-tested after changes. Several studies demonstrate the potential of these algorithms to identify optimal test sets, streamlining the testing process.

Quantum machine learning represents another key area. Researchers explore the use of quantum neural networks (QNNs) and Quantum Extreme Learning Machines (QELMs) for tasks such as vulnerability detection and quality of service (QoS) prediction. QNNs, computational models inspired by biological neural networks but leveraging quantum phenomena, show promise in identifying security flaws and predicting system performance, potentially exceeding the capabilities of classical machine learning models. Applications extend to diverse systems, including cancer registries and elevator control systems, indicating a move towards practical implementation.

Beyond testing and machine learning, researchers investigate quantum-enhanced software analysis. Quantum algorithms are utilised to improve static analysis techniques – methods for evaluating code without executing it – and identify code clones, enhancing software quality assurance by leveraging the computational power of quantum systems.

QBSE actively seeks to re-formulate software engineering problems to intrinsically benefit from quantum computational properties, rather than merely accelerating existing classical solutions. This guides research towards identifying problem types where quantum optimisation, search, and learning techniques offer a fundamental advantage. Despite increasing research activity, the field remains fragmented, spanning a range of applications and algorithms, lacking a cohesive framework. The prevalence of pre-print publications highlights the rapid evolution of the field, but also underscores the need for rigorous peer review and validation of reported results.

Future research should prioritise establishing a structured research agenda for QBSE, including defining clear metrics for evaluating the performance of quantum algorithms in software engineering contexts, and conducting comparative analyses against established classical techniques. Further investigation into the scalability and practical limitations of current approaches is also crucial. While quantum annealing and QAOA currently dominate, investigating alternative quantum algorithms and hybrid quantum-classical approaches may unlock further performance gains. Fostering collaboration between quantum computing experts and software engineering practitioners will be essential to translate theoretical advancements into practical, real-world applications.