We consider a family of unstructured problems, for which we propose a method for constructing analog, continuous-time quantum algorithms that are more efficient than their classical counterparts. In this family of problems, which we refer to as `scrambled output' problems, one has to find a minimum-cost configuration of a given integer-valued n-bit function whose output values have been scrambled in some arbitrary way. Special cases within this set of problems are Grover's search problem of finding a marked item in an unstructured database, certain random energy models, and the functions of the Deutsch-Josza problem. We consider a couple of examples in detail. In the first, we provide a deterministic analog quantum algorithm to solve the seminal problem of Deutsch and Josza, in which one has to determine whether an n-bit boolean function is constant (gives 0 on all inputs or 1 on all inputs) or balanced (returns 0 on half the input states and 1 on the other half). We also study one variant of the random energy model, and show that, as one might expect, its minimum energy configuration can be found quadratically faster with a quantum adiabatic algorithm than with classical algorithms.
在黑盒模型中,约束的问题的承诺是唯一的承认量子指数加速超过最好的经典算法的查询复杂性。最著名的例子是Deutsch-Jozsa算法。最近,Wim van Dam提出了一种用于非结构化问题的算法(即,没有承诺的人)。我们认为Deutsch-Jozsa算法的限制性较低(或“破碎”)的承诺,并研究过渡到一个非结构化的问题。我们将其与范达姆算法的成功进行比较。这两个比较与一个标准的经典采样算法。Deutsch-Jozsa算法仍然很好,因为问题最初变得不那么结构化,但van Dam算法可以适应,以便在承诺减弱时变得优于Deutsch-Jozsa算法。
Achieving double-logarithmic precision dependence in optimization-based quantum unstructured search
Beijing International Center for Mathematical Research, Peking University(北京大学数学中心);Yau Mathematical Sciences Center, Tsinghua University(清华大学丘成桐数学科学中心)