Near-Optimal Fidelity in Quantum Circuits through Incorporating Efficient Real-time Error Based Heuristics in Compiler Mappings

TL;DR

This paper introduces a heuristic quantum circuit mapping method that incorporates real-time error feedback and device connectivity, significantly improving fidelity over existing algorithms.

Contribution

The paper presents a novel heuristic technique, CAES, that effectively uses real-time error data and connectivity info to enhance quantum circuit mapping fidelity.

Findings

CAES outperforms baseline algorithms by 1.62x and 1.934x on average.

CAES achieves 1.7x better success rate than the state-of-the-art.

Extensive experiments validate the effectiveness of the proposed heuristics.

Abstract

To run a quantum program in the real device, the compiler maps the logical qubits to physical qubits. This is the most crucial step of compiling a quantum circuit. Because the fidelity of a quantum circuit depends heavily on this mapping process. However, this qubit mapping problem is NP-complete. Therefore, we should resort to heuristics to find high-fidelity mappings. In this paper, we focused on finding efficient heuristic techniques to incorporate real-time error feedback and device connectivity information in order to achieve high fidelity mapping of the quantum circuits. We performed extensive analysis and experimental study based on two baseline algorithms. We performed our experimentation on various combinations of different error rates and heuristic techniques. Consequently, we designed very elegant techniques to consider both all types of real-time error feedback and connectivity information. We showed that our best heuristic approach performs \textbf{1.62x} ( on average) better than one baseline and \textbf{1.934x} ( on average ) better than the other baseline on random benchmarks. Finally, we compared our best heuristic ( CAES ) with the state-of-the-art heuristic-based mapping algorithm on representative benchmarks. We found that CAES performed \textbf{1.7x} ( on average ) better than the state of the art in terms of success rate.