Quantum Approximate Counting with Nonadaptive Grover Iterations

TL;DR

This paper investigates quantum approximate counting using nonadaptive Grover iterations, demonstrating that such algorithms can achieve optimal query complexity without adaptivity, simplifying implementation.

Contribution

It introduces a nonadaptive Grover iteration algorithm for approximate counting, matching the optimal query complexity of adaptive methods.

Findings

Nonadaptive Grover iteration algorithms achieve $O(rac{ ext{sqrt}(N)}{ ext{epsilon}})$ query complexity.

The nonadaptive approach is proven to be asymptotically optimal.

Simplifies quantum approximate counting by removing adaptivity requirements.

Abstract

Approximate Counting refers to the problem where we are given query access to a function $f : [N] \to \{0,1\}$, and we wish to estimate $K = #\{x : f(x) = 1\}$ to within a factor of $1+\epsilon$ (with high probability), while minimizing the number of queries. In the quantum setting, Approximate Counting can be done with $O\left(\min\left(\sqrt{N/\epsilon}, \sqrt{N/K}/\epsilon\right)\right)$ queries. It has recently been shown that this can be achieved by a simple algorithm that only uses "Grover iterations"; however the algorithm performs these iterations adaptively. Motivated by concerns of computational simplicity, we consider algorithms that use Grover iterations with limited adaptivity. We show that algorithms using only nonadaptive Grover iterations can achieve $O\left(\sqrt{N/\epsilon}\right)$ query complexity, which is tight.