Researchers Raphaël Bouchet and Alena Puchkova have developed and evaluated a proof-of-concept quantum clustering framework designed to segment customer groups in supply chains. The investigation, published in Scientific Reports, examined whether quantum annealing can practically enhance supply chain customer segmentation compared with classical K-Means algorithms. The researchers translated the multi-cluster data partitioning challenge into a Quadratic Unconstrained Binary Optimisation formulation to systematically test alternative optimization mechanisms against established industry standards.
To benchmark performance, the authors tested customer datasets ranging in scale from 200 to 5,000 data points using classical Simulated Annealing alongside D-Wave’s Leap hybrid solver. The research team introduced a dedicated penalty parameter optimisation pipeline specifically to address known solver sensitivity during the optimization process. The team systematically benchmarked execution runtime, accuracy metrics, and scalability across these dataset sizes to determine current deployment feasibility.
The comparative benchmarking revealed distinct operational trade-offs between the classical and quantum computing approaches. Classical K-Means maintained superior cluster stability and lower inertia across the evaluated benchmark datasets. In contrast, the hybrid quantum configuration delivered faster solver-reported runtimes across the tests. However, severe solver sensitivity required meticulous penalty parameter tuning before the quantum formulation could produce competitive mathematical clustering solutions across the testing conditions.
On a simplified benchmark featuring 200 data points across two features, the hybrid quantum model showed clear metric advantages once penalties were properly optimised. Under these specific conditions, it delivered improved silhouette, Davies-Bouldin, Calinski-Harabasz, and Dunn scores relative to K-Means. The authors noted that these evaluation metrics inherently reward balanced partitions, which are structurally favoured by the mathematical formulation of the pairwise QUBO objective.
Despite these localized metric improvements, the researchers concluded that classical clustering methods remain dominant for industrial-scale supply chain segmentation. Practical deployment beyond mid-sized datasets remains strictly constrained by current qubit capacity and solver access limitations. Nevertheless, the demonstrated QUBO framework, penalty-optimisation pipeline, and objective-level analysis establish an essential documented foundation for future quantum-assisted supply chain clustering as hardware technology matures.
Subject of Research: Quantum annealing for supply chain customer segmentation
Article Title: Quantum annealing for customer segmentation: a proof-of-concept QUBO framework for supply chain clustering
Article References: Bouchet, R., & Puchkova, A. (2026). Quantum annealing for customer segmentation: a proof-of-concept QUBO framework for supply chain clustering. Scientific Reports. https://doi.org/10.1038/s41598-026-75413-3
Image Credits: AI Generated
DOI: 10.1038/s41598-026-75413-3
Keywords: quantum annealing, customer segmentation, supply chain clustering, QUBO formulation, K-Means, Simulated Annealing, D-Wave Leap, hybrid solver, penalty parameter optimisation, silhouette score, Davies-Bouldin index, Calinski-Harabasz score, Dunn index, qubit capacity, cluster inertia, balanced partitions, runtime benchmarking, quadratic unconstrained binary optimisation
Cite Scienmag News
Katie Riggs. (November 19, 2026). Quantum Annealing Framework Benchmarked for Supply Chain Customer Segmentation. Scienmag. https://scienmag.com/quantum-annealing-framework-benchmarked-for-supply-chain-customer-segmentation/
Katie Riggs. "Quantum Annealing Framework Benchmarked for Supply Chain Customer Segmentation." Scienmag, 19 November 2026, https://scienmag.com/quantum-annealing-framework-benchmarked-for-supply-chain-customer-segmentation/. Accessed 20 November 2026.
Katie Riggs. "Quantum Annealing Framework Benchmarked for Supply Chain Customer Segmentation." Scienmag. November 19, 2026. https://scienmag.com/quantum-annealing-framework-benchmarked-for-supply-chain-customer-segmentation/








