| Title |
Research of quantum computer topologies for optimization algorithms |
| Translation of Title |
Kvantinių kompiuterių topologijų tyrimas optimizavimo algoritmams. |
| Authors |
Keibas, Gabrielius |
| Full Text |
|
| Pages |
120 |
| Keywords [eng] |
kvantinių schemų optimizavimas, kubitų atvaizdavimo problema, kvantiniai optimizavimo algoritmai, kvantinių kompiuterių topologijos, quantum circuit mapping, qubit mapping problem, quantum optimization algorithms, quantum computer topologies |
| Abstract [eng] |
Efficient execution of quantum optimisation algorithms on near-term quantum hardware remains challenging due to limited qubit connectivity. This work presents QuMeld, an open-source framework for topology-aware optimisation of quantum circuits. It is used to systematically compare six qubit mapping algorithms across sixteen quantum computer topologies and six benchmark circuits covering VQE and QAOA workloads. Results reveal that quantum optimisation algorithms do not reduce the qubit mapping problem to a simplified form. Preliminary decision tree rules achieved 45,7 – 72,3 % accuracy, outperforming the random baseline (16,7 %). While this exceeds random selection, it is insufficient to replace exhaustive mapper evaluation in practice. The framework was validated on real IBM hardware with results that demonstrated end-to-end hardware compatibility, and a quantum machine learning experiment on a medical benchmark dataset showed practical end-to-end usability. Overall, no single mapper is globally optimal, including default framework mappers, and criterion-driven exhaustive evaluation across candidate mappers is the recommended practical strategy in the current NISQ era. |
| Dissertation Institution |
Vilniaus universitetas. |
| Type |
Master thesis |
| Language |
English |
| Publication date |
2026 |
