量子电子学报 ›› 2021, Vol. 38 ›› Issue (1): 75-85.

• 量子光学 • 上一篇    下一篇

基于预评价的量子电路线性最近邻综合算法

王艺臻1, 管致锦1;2∗, 管海宇1   

  1. 1 南通大学信息科学技术学院, 江苏南通226019; 2 江苏省专用集成电路重点实验室, 江苏南通226019
  • 收稿日期:2019-12-26 修回日期:2020-02-14 出版日期:2021-01-28 发布日期:2021-02-01
  • 通讯作者: E-mail: guan.zj@ntu.edu.cn E-mail:guan.zj@ntu.edu.cn
  • 作者简介:王艺臻( 1996 - ), 女, 江苏南通人, 研究生, 主要从事量子电路综合与二维量子电路近邻化方面的研究。 E-mail: W−yz1996@hotmail.com
  • 基金资助:
    Supported by National Natural Science Foundation of China (国家自然科学基金, 62072259), Natural Science Foundation of Jiangsu Province (江苏省自然科学基金, BK20151274), Postgraduate Research and Practice Innovation Program of Jiangsu Province (江苏省研究生科研与实践 创新计划项目, KYCX17−1916)

Linear nearest neighbor synthesis algorithm of quantum circuits based on pre-evaluation

WANG Yizhen1, GUAN Zhijin1;2∗, GUAN Haiyu1   

  1. 1 College of Information Science and Technology, Nantong University, Nantong 226019, China; 2 Jiangsu Key Laboratory of Asic Design, Nantong University, Nantong 226019, China
  • Received:2019-12-26 Revised:2020-02-14 Published:2021-01-28 Online:2021-02-01

摘要: 线性最近邻(LNN) 量子电路是一种重要的量子电路架构, 它为量子电路的物理实现奠定了基础。为了 构造LNN 量子逻辑电路, 提出了一种基于预评价的量子电路线性最近邻逻辑综合算法。该算法首先对部分非 近邻量子电路进行全局换线; 接着启发式地对换线后的量子电路进行评价, 通过求解最小混乱值找到最近邻代 价最小的量子电路集; 最后添加交换门, 以最少的量子代价将该量子电路转换为最近邻结构。该算法通过预评 价, 一方面可以准确计算出量子电路中是否具有可删除的冗余交换门并删除之; 另一方面使每一个非近邻量子 门转化为LNN 量子门时添加交换门的数量最小, 进而得到量子代价最小的量子电路。实验结果表明, 与已知算 法相比, 该算法能以更低的量子电路平均门数、更高的量子代价优化率适用于更广范围内的量子电路近邻化。

关键词: 量子电路, 线性最近邻, 交换门, 量子代价

Abstract: Linear nearest neighbor (LNN) quantum circuit is an important quantum circuit architecture, which lays a foundation for the physical realization of quantum circuit. In order to construct LNN quantum logic circuits, a linear nearest neighbor logic synthesis algorithm based on pre-evaluation is proposed. In this algorithm, partial non-nearest neighbor quantum circuits are globally reordered firstly, then the reordered quantum circuits are evaluated by heuristic method, and the set of quantum circuits with the lowest nearest neighbor cost is found by computing the minimum chaos value. Finally, SWAP gates are added to convert the quantum circuit to a nearest neighbor structure with the lowest quantum cost. Through pre-evaluation, on the one hand, the algorithm can accurately calculate whether there are redundant SWAP gates that can be deleted in the quantum circuit and delete them. On the other hand, it can minimize the number of SWAP gates added when each non-nearst neighbor quantum gate is converted into LNN quantum gate, and then the quantum circuit with the lowest quantum cost is obtained. The experimental results show that compared with the known algorithm, the proposed algorithm has less average gates, higher optimization rate of quantum cost and a wider application range.

Key words: quantum circuit, linear nearest neighbor, SWAP gate, quantum cost

中图分类号: