AQUA: an Agile Process to Develop Quantum Annealing Applications

TL;DR

AQUA introduces an agile, milestone-driven software engineering process tailored to quadratically constrained binary optimization problems solved via quantum annealing and related QA approaches. Built on a Design Science Research foundation, it combines Scrum-like iteration with four phased gates (initial assessment, prototype strategy, system development, deployment/maintenance) to manage the unique math and hardware constraints of QA/DQA projects. The approach is demonstrated through a credit-scoring case, showing how problem formulation, embedding choices, and hybrid classical–quantum workflows can be structured from the problem charter to deployment. The study argues that AQUA provides a practical blueprint for organizations to adopt QA technologies, while identifying needs for quantitative milestones, embedding heuristics, and broader validation across domains.

Abstract

Quadratic unconstrained binary optimization (QUBO) is a field of operations research that is attracting growing interest due to the recent availability of quantum hardware targeted at solving QUBO problems. However, practical adoption is hindered by mathematical intricacy, hardware constraints, and a lack of sound software engineering processes for QUBO development. This work presents AQUA (Agile QUantum Annealing), an agile lifecycle for QUBO/QA development created through an industry-academia partnership between NetService S.p.A and the University of Cagliari. Using the Design Science Research (DSR) approach, AQUA customizes Scrum to the needs of QUBO/QA development, structuring work into four stages: initial assessment with formal modeling, prototype-driven algorithm selection, agile implementation, and deployment with ongoing maintenance, each gated by milestones. Validated on a real credit-scoring case, AQUA shows feasibility and offers an explicit, systematic QA engineering framework. Key contributions of our work are: a dedicated QUBO/QA software process, its creation and design using DSR approach, and its empirical validation on a simple yet real case study.

Figures (4)

• Figure 1: Design Research outline: activities, methodologies and outputs.
• Figure 2: An overview of AQUA process.
• Figure 3: The steps of the "Initial Assessment" and “Computational strategy and software design" phases.
• Figure 4: The steps and decisions of the "System Development" and "System validation and maintenance" phases.