F. Gemeinhardt, S. Klikovits, M. Wimmer: GeQuPI: Quantum Program Improvement with Multi-Objective Genetic Programming, Software Engineering 2025 Companion Proceedings, Fachtagung des GI-Fachbereichs Softwaretechnik, Karlsruhe, Germany, February 24-28, 2025. Doi: 10.18420/se2025-ws-19
To tackle the challenge of debugging and optimizing quantum programs, we propose the Genetic Quantum Program Improver (GeQuPI), a framework for the automated improvement of the accuracy and efficiency of quantum circuits. Our framework uses techniques from quantum information theory and applies multi-objective genetic programming, which can be further hybridized with quantum-aware optimizers. To evaluate GeQuPI, it is applied to 47 quantum programs reused from literature and publicly available libraries. The results of the evaluation demonstrate that GeQuPI is capable of correcting faulty programs and optimizing inefficient ones for a large share of the studied programs, showing an average improvement of efficiency of 35%.