Generalized quantum optimization algorithms for near-term quantum computing architectures

用于近期量子计算架构的广义量子优化算法

• 批准号: 2327797
• 金额: --
• 依托单位: University College London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2019
• 资助国家: 英国
• 起止时间: 2019 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2327797
• 关键词: Generalized; quantum; optimization; algorithms; near

In a distant future, large-scale, ideal quantum computers will speed up many interesting computational tasks by leveraging a restricted number of fundamental quantum algorithms. However, the rst generation of quantum computers |those that exist now or that we may see in a few years| will be error-prone, hence merely capable of running a restricted set of quantum programmes. In this context, it is pressing to find valuable applications for these machines. Obtaining good solutions to hard optimization problems could be such an application.The Quantum Approximate Optimization Algorithm (QAOA) is the best-established quantum optimization algorithm for near-term hardware to date. However, as of today, the advantage of QAOA over classical algorithms has remained elusive. More specifically, a series of recent results established significant limitations on the most naive versions of QAOA - though the relevance of these results to the intermediate-scale problems addressable by near-term quantum computers may be questioned.Starting from the present understanding of the limitations of QAOA, the project will aim at designing generalized versions of the algorithm and assess their performance. Potential applications of these building blocks as shallow approximations of common quantum algorithms will then be explored. Several applications of these approximate algorithms will subsequently be considered, possibly including the enhancement of classical optimization heuristics or algorithms from the variational quantum eigensolver family |which are central to quantum chemistry or material science simulations on a quantum computer.

在遥远的未来，大规模的理想量子计算机将通过利用有限数量的基本量子算法来加速许多有趣的计算任务。然而，第一代量子计算机|那些现在存在的或我们几年后可能会看到的|容易出错，因此只能运行一组有限的量子程序。在这种情况下，迫切需要为这些机器找到有价值的应用。量子近似优化算法（QAOA）是迄今为止针对近期硬件的最佳量子优化算法。然而，到今天为止，QAOA相对于经典算法的优势仍然难以捉摸。更具体地说，最近的一系列结果对QAOA的最幼稚版本建立了显着的限制-尽管这些结果与近期量子计算机可解决的中等规模问题的相关性可能受到质疑。从目前对QAOA限制的理解开始，该项目将旨在设计该算法的通用版本并评估其性能。然后将探索这些构建块作为常见量子算法的浅近似的潜在应用。随后将考虑这些近似算法的几个应用，可能包括经典优化算法的增强或来自变分量子本征解算器族的算法|这对量子计算机上的量子化学或材料科学模拟至关重要。