We consider the problem of transforming a given symmetric matrix as close to block diagonal as possible by symmetric permutations of its rows and columns. Such a problem arises, for example, as a preprocessing for the block Jacobi method for the symmetric eigenvalue problem. To solve this problem on a quantum annealing machine, Teramoto et al. proposed its QUBO (Quadratic Unconstrained Binary Optimization) formulation and strategies for embedding the resulting QUBO into the Pegasus network of the D-Wave Advantage quantum annealer. In this paper, we propose two more embedding strategies based on the minimum-cost flow and integer multi-commodity flow. We also propose an alternative QUBO formulation for which the connectivity graph has a more local structure. Numerical experiments show that our new QUBO formulation requires less physical qubits when embedding the problem into D-Wave Advantage’s Pegasus network.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Comparison of Two QUBO Formulations of Approximate Block Diagonalization and Their Performance on the D-Wave Advantage Quantum Annealing Machine

  • Koushi Teramoto,
  • Shuhei Kudo,
  • Yusaku Yamamoto

摘要

We consider the problem of transforming a given symmetric matrix as close to block diagonal as possible by symmetric permutations of its rows and columns. Such a problem arises, for example, as a preprocessing for the block Jacobi method for the symmetric eigenvalue problem. To solve this problem on a quantum annealing machine, Teramoto et al. proposed its QUBO (Quadratic Unconstrained Binary Optimization) formulation and strategies for embedding the resulting QUBO into the Pegasus network of the D-Wave Advantage quantum annealer. In this paper, we propose two more embedding strategies based on the minimum-cost flow and integer multi-commodity flow. We also propose an alternative QUBO formulation for which the connectivity graph has a more local structure. Numerical experiments show that our new QUBO formulation requires less physical qubits when embedding the problem into D-Wave Advantage’s Pegasus network.