arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2411.00518quant-ph
收藏

量子树生成器提升knapsack问题的QAOA前沿水平

Quantum tree generator improves QAOA state-of-the-art for the knapsack problem

d-fine GmbH(d-fine有限公司);Institut für Theoretische Physik Leibniz Universität Hannover(汉诺威莱布尼茨大学理论物理研究所)

Paul Christiansen, Lennart Binkowski, Debora Ramacciotti, Sören Wilkening

AI总结:

量子树生成器提升QAOA在knapsack问题中的性能,通过结合Grover-mixer框架提出AAM-QAOA,实现更优解采样。

摘要:

本文介绍了一种针对knapsack问题的量子近似优化算法(QAOA)的新方法。我们结合最近提出的量子树生成器作为所有可行解的高效状态准备电路,与Grover-mixer QAOA框架结合,形成首个Amplitude Amplification-mixer QAOA(AAM-QAOA)代表。在包含多达20个knapsack物品的困难基准集中,我们展示了该方法在当前前沿Copula-QAOA上的改进性能。然而,对于较大的实例规模,两种方法都无法在降序价值-重量比贪心打包策略之上提供更好的结果,至少在考虑的电路深度范围内。然而,对于足够高的电路深度,我们能够证明AAM-QAOA最终能够采样到最优解。

英文摘要:

This paper introduces a novel approach to the Quantum Approximate Optimization Algorithm (QAOA), specifically tailored to the knapsack problem. We combine the recently proposed quantum tree generator as an efficient state preparation circuit for all feasible solutions to the knapsack problem with the framework of Grover-mixer QAOA to form the first representative of Amplitude Amplification-mixer QAOA (AAM-QAOA). On hard benchmark sets with up to 20 knapsack items, we demonstrate our method's improved performance over the current state-of-the-art Copula-QAOA. However, for larger instance sizes, both approaches fail to deliver better outcomes than greedily packing items in descending value-to-weight ratio, at least for the considered circuit depths. For sufficiently high circuit depths, however, we can prove that AAM-QAOA will eventually be able to sample the optimal solution.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑