Quantum computers are becoming a reality today due to the rapid progress made by researchers in the last years. In the process of building quantum computers, IBM has developed several versions—starting from 5-qubit architectures like IBM QX2 and IBM QX4 to larger 16- or 20-qubit architectures. These architectures support arbitrary rotations of a single qubit and a controlled negation (CNOT) involving two qubits. The two qubit operations come with added coupling-map restrictions that only allow specific physical qubits to be the control and target qubits of the operation. In order to execute a quantum circuit on the IBM QX architecture, CNOT gates must satisfy the so-called coupling constraints of the architecture. Previous works addressed this issue with the objective of reducing the number of gates and the circuit depth. However, in this article, we show that further improvements are possible. To this end, we present a general approach for further improving the number of gate operations and depth of the mapped circuit. The proposed approach encompasses the selection of physical qubits, determining initial and local permutations efficiently to obtain the final circuit mapped to the given IBM QX architecture. Through experiments, improvements are observed over existing methods in terms of the number of gates and circuit depth.

中文翻译：

---

改进了量子电路到 IBM QX 架构的映射

由于研究人员在过去几年取得的快速进步，量子计算机正在成为现实。在构建量子计算机的过程中，IBM 开发了多个版本——从 IBM QX2 和 IBM QX4 等 5 量子位架构到更大的 16 或 20 量子位架构。这些架构支持单个量子位的任意旋转和涉及两个量子位的受控否定 (CNOT)。这两个量子位操作带有附加的耦合图限制，仅允许特定的物理量子位作为操作的控制和目标量子位。为了在 IBM QX 架构上执行量子电路，CNOT 门必须满足架构的所谓耦合约束。以前的工作解决了这个问题，目的是减少门的数量和电路深度。然而，在这篇文章中，我们表明进一步改进是可能的。为此，我们提出了一种进一步提高映射电路的门操作数量和深度的通用方法。所提出的方法包括物理量子位的选择，有效地确定初始和局部排列以获得映射到给定 IBM QX 架构的最终电路。通过实验，观察到在门的数量和电路深度方面优于现有方法的改进。