Resource-aware QAOA for autoclave batch scheduling in composite manufacturing using QUBO decomposition and parameter transfer

Abstract Industrial quantum computing requires problem decompositions that respect near-term resource limits while preserving operational value. This study presents a hybrid classical-quantum workflow for autoclave batch scheduling in composite manufacturing, in which release-time and cure-family logic remain in a classical orchestration layer and each batch-selection decision is encoded as a quadratic unconstrained binary optimization (QUBO) problem. Candidate pools of 4–10 jobs produce 8–14 logical-qubit QAOA subproblems after capacity-slack encoding. Depth-1 and depth-2 QAOA were evaluated using noiseless statevector simulation, with depth-2 parameters trained on four instances per problem size and transferred without re-optimization to six held-out instances per size. For transferred depth-2 QAOA, decoded feasible probability mass decreased from 0.956 at 8 qubits to 0.427 at 14 qubits, while optimal-subset probability decreased from 0.05283 to 0.00150. Comparison with an exact uniform-sampling null showed no consistent optimal-subset concentration advantage for transferred $$\text{p}=2$$ QAOA, despite exact best-of-4096 recovery across the held-out instances. The estimated sampling budget for a 95% probability of observing the optimum increased from approximately 55 to 2000 shots. In 15 end-to-end synthetic production scenarios containing 24–48 jobs, no statistically significant differences were detected between transferred QAOA and exact per-batch selection or simulated annealing for any reported schedule metric. Relative to greedy selection, transferred QAOA increased mean autoclave utilization by 2.70% points, corresponding to a 3.3% relative increase. This was the only nominally significant paired difference against greedy selection (unadjusted Wilcoxon $$\:p=0.029$$ ); other schedule metrics improved directionally but were not statistically significant. All results were obtained using noiseless statevector simulation and establish neither quantum advantage nor hardware-level speedup, but rather a testable quantum-classical integration pathway defined by decomposition, parameter reuse, sampling burden, and classical fallback.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-12
DOI
https://doi.org/10.1038/s41598-026-75912-3
Primary Topic
Quantum Computing Algorithms and Architecture
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Resource-aware QAOA for autoclave batch scheduling in composite manufacturing using QUBO decomposition and parameter transfer

Obinna Onyebuchi Barah, Ige Bori, Zaid Abdulhamid Alhulaybi Albin Zaid, Khalad A. AlMuhaysh et al.
1 links
Scientific Reports
Quantum Computing Algorithms and Architecture
article

Resource-aware QAOA for autoclave batch scheduling in composite manufacturing using QUBO decomposition and parameter transfer

Obinna Onyebuchi Barah, Ige Bori, Zaid Abdulhamid Alhulaybi Albin Zaid, Khalad A. AlMuhaysh, Abdulrazak Jinadu Otaru
article en

Abstract

Abstract Industrial quantum computing requires problem decompositions that respect near-term resource limits while preserving operational value. This study presents a hybrid classical-quantum workflow for autoclave batch scheduling in composite manufacturing, in which release-time and cure-family logic remain in a classical orchestration layer and each batch-selection decision is encoded as a quadratic unconstrained binary optimization (QUBO) problem. Candidate pools of 4–10 jobs produce 8–14 logical-qubit QAOA subproblems after capacity-slack encoding. Depth-1 and depth-2 QAOA were evaluated using noiseless statevector simulation, with depth-2 parameters trained on four instances per problem size and transferred without re-optimization to six held-out instances per size. For transferred depth-2 QAOA, decoded feasible probability mass decreased from 0.956 at 8 qubits to 0.427 at 14 qubits, while optimal-subset probability decreased from 0.05283 to 0.00150. Comparison with an exact uniform-sampling null showed no consistent optimal-subset concentration advantage for transferred $$\text{p}=2$$ QAOA, despite exact best-of-4096 recovery across the held-out instances. The estimated sampling budget for a 95% probability of observing the optimum increased from approximately 55 to 2000 shots. In 15 end-to-end synthetic production scenarios containing 24–48 jobs, no statistically significant differences were detected between transferred QAOA and exact per-batch selection or simulated annealing for any reported schedule metric. Relative to greedy selection, transferred QAOA increased mean autoclave utilization by 2.70% points, corresponding to a 3.3% relative increase. This was the only nominally significant paired difference against greedy selection (unadjusted Wilcoxon $$\:p=0.029$$ ); other schedule metrics improved directionally but were not statistically significant. All results were obtained using noiseless statevector simulation and establish neither quantum advantage nor hardware-level speedup, but rather a testable quantum-classical integration pathway defined by decomposition, parameter reuse, sampling burden, and classical fallback.

Scientific ReportsVol. 16(1)
Kampala International University (UG), King Faisal University (SA), Federal University of Technology Minna (NG)
King Faisal University
Quantum Computing Algorithms and Architecture
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon