Exploring the Solution of Quadratic Unconstrained Binary Optimization Problems with Different Quantum Hardware¶
Qifeng Pan (HLRS)
Abstract¶
In my previous WSSP talk, I introduced the Quadratic Unconstrained Binary Optimization (QUBO) problem, a fundamental classical optimization problem that is NP-hard. Researchers have been investigating whether quantum algorithms can provide a polynomial-time advantage for solving such problems. Last time, I presented the LR-QAOA algorithm, which offers better scalability compared to the standard QAOA. However, as the number of qubits increases, LR-QAOA requires deeper circuits to maintain its ability to find optimal solutions. Experiments show that when the circuit depth is increased linearly, the success rate of LR-QAOA decreases dramatically in test cases. In this talk, I will introduce improvements to the LR-QAOA method, including polynomial and exponential scheduling schemes as well as gauge operators. Experiments clearly demonstrate that these methods significantly improve the success rate of LR-QAOA. In addition, I will present performance comparisons across different quantum hardware backends, including photonic quantum computer (QC) from OCRA, NV-center QC from Quantum Brilliance, and superconducting QC from IBM. Finally, I will discuss the scaling potential of these platforms in comparison to classical methods.