Feedback-assisted quantum search by continuous-time quantum walks

TL;DR

This paper introduces a feedback-assisted quantum search method using continuous-time quantum walks on a cycle graph, employing measurement and dynamic feedback to efficiently locate a target node, with performance analyzed under various control constraints.

Contribution

It presents a novel feedback-based quantum search protocol with a dynamical oracle, demonstrating its effectiveness through numerical simulations and control constraint analysis.

Findings

Protocol quickly localizes the walker on the target node with unbounded controls.

Performance decreases with bounded control amplitudes but still stabilizes the walker.

Discrete feedback protocols are nearly as effective as continuous ones.

Abstract

We address the quantum search of a target node on a cycle graph by means of a quantum walk assisted by continuous measurement and feedback. Unlike previous spatial search approaches, where the oracle is described as a projector on the target state, we instead consider a dynamical oracle implemented through a feedback Hamiltonian. In particular, our protocol is able to drive the walker to a desired target node. The idea is based on continuously monitoring the position of the quantum walker on the graph and then to apply a unitary feedback operation based on the information obtained from measurement. The feedback changes the couplings between the nodes and it is optimized at each time via a numerical procedure. We numerically simulate the stochastic trajectories describing the evolution for graphs of dimensions up to $N=15$, and quantify the performance of the protocol via the average fidelity between the state of the walker and the target node. We discuss different constraints on the control strategy, in particular on the possible values that the feedback couplings can take. We find that for unbounded controls, the protocol is able to quickly localize the walker on the target node. We then discuss how the performance is lowered by posing an upper bound on the control couplings, but still allowing to stabilize the walker in the desired node. Finally, we show how a digital feedback protocol, where the couplings can take values only from a discrete set, seems in general as efficient as the continuous bounded one.