Experimental implementation of quantum greedy optimization on quantum computer

TL;DR

This paper demonstrates an improved quantum greedy optimization algorithm implemented on a quantum computer, which reduces the number of measurements needed to find ground states by enhancing sensitivity analysis techniques.

Contribution

It introduces an improved sensitivity analysis method for d-QGO that employs larger differential intervals, making the algorithm more practical on noisy quantum devices.

Findings

Reduces shot count for sensitivity analysis

Maintains high success probability in ground state determination

Enhances practicality of quantum greedy optimization

Abstract

This paper implements a quantum greedy optimization algorithm based on the discretization of time evolution (d-QGO). Quantum greedy optimization, which was originally developed for reducing processing time via counterdiabatic driving, sequentially selects a parameter in the counterdiabatic term from the sensitivity analysis of energy and then determines the parameter value. For implementing d-QGO on a quantum computer, the sensitivity analysis may become a bottleneck to find the ground state in a short time due to device and shot noise. In this paper, we present an improved sensitivity analysis for d-QGO that employs a sufficiently large differential interval. We demonstrate that d-QGO reduces the number of shots required to determine the sensitivity while maintaining the success probability.