Genetic algorithm for searching bipolar Single-Flux-Quantum pulse sequences for qubit control

TL;DR

This paper introduces a genetic algorithm to optimize superconducting digital control sequences using SFQ pulses for qubit operations, reducing leakage and memory usage, demonstrated with high-fidelity single-qubit gates.

Contribution

It presents a novel genetic algorithm for designing SFQ pulse sequences that minimize qubit leakage and optimize memory, adaptable to various system parameters.

Findings

Achieved over 99.99% fidelity for a π/2 rotation gate.

Algorithm finds optimized sequences within reasonable time.

Potential application to multi-qubit systems in future work.

Abstract

Nowadays most of superconducting quantum processors use charge qubits of a transmon type. They require implementation of energy efficient qubit state control scheme. A promising approach is the use of superconducting digital circuits operating with single-flux-quantum (SFQ) pulses. The duration of SFQ pulse control sequence is typically larger than that of conventional microwave drive pulses but its length can be optimized for the system with known parameters. Here we introduce a genetic algorithm for unipolar or bipolar SFQ control sequence search that minimize qubit state leakage from the computational subspace. The algorithm is also able to find a solution in the form of a repeating subsequence in order to save memory on the control chip. Its parallel implementation can find the appropriate sequence for arbitrary system parameters from a practical range in a reasonable time. The algorithm is illustrated by the example of the rotation gate around the axis by an angle $\pi/2$ with fidelity over 99.99%. In this paper, we present the results for a single-qubit system, but in the future we will apply the developed approach to study a system of two qubits.