Scholarlink

<p>We examine the significant potential of quantum computing for solving combinatorial optimization problems in supply chains, with a focus on recent advances and challenges that lie ahead. We observe that the constructs in classical quantum physics pose three main gaps for exploiting quantum advantage in the supply chain area: optimization modeling, algorithm design, and solution assessment. To overcome these gaps, we elaborate how the Quadratic Unconstrained Binary Optimization (QUBO) model that lies at the heart of supply chain optimization with quantum computing can be fruitfully exploited within the framework of Quantum Bridge Analytics (QBA) to connect quantum computing and classical optimization. Our QBA/QUBO approach is implemented for the well-known uncapacitated facility location problem (UFLP) in supply chain network design, using Gurobi’s QUBO solver, D-Wave’s hybrid classical-quantum solver, and the proprietary next generation quantum (NGQ) solver. A comprehensive computational study is performed on the UFLP benchmark instances with comparisons to four recent algorithms for UFLP in the literature. The QBA/QUBO approach achieved high-quality results for problems of all sizes, solving all the small-size and some medium-size instances with up to 10,000 decision variables to optimality in less than 20&#xa0;s, and achieving less than 3% optimality gap for large instances with more than 250,000 decision variables in less than 10&#xa0;min on average. We also find that in addition to problem size, the polyhedral properties of a UFLP instance may significantly impact the efficiency of quantum annealing based algorithm.</p>

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

Quantum bridge analytics: a QUBO approach to the uncapacitated facility location problem

  • 期刊论文

  • 发表日期：

    2026年4月9日

查看全文
  • Haitao Li,
  • Rick Hennig,
  • Gary Kochenberger,
  • Fred Glover

摘要

We examine the significant potential of quantum computing for solving combinatorial optimization problems in supply chains, with a focus on recent advances and challenges that lie ahead. We observe that the constructs in classical quantum physics pose three main gaps for exploiting quantum advantage in the supply chain area: optimization modeling, algorithm design, and solution assessment. To overcome these gaps, we elaborate how the Quadratic Unconstrained Binary Optimization (QUBO) model that lies at the heart of supply chain optimization with quantum computing can be fruitfully exploited within the framework of Quantum Bridge Analytics (QBA) to connect quantum computing and classical optimization. Our QBA/QUBO approach is implemented for the well-known uncapacitated facility location problem (UFLP) in supply chain network design, using Gurobi’s QUBO solver, D-Wave’s hybrid classical-quantum solver, and the proprietary next generation quantum (NGQ) solver. A comprehensive computational study is performed on the UFLP benchmark instances with comparisons to four recent algorithms for UFLP in the literature. The QBA/QUBO approach achieved high-quality results for problems of all sizes, solving all the small-size and some medium-size instances with up to 10,000 decision variables to optimality in less than 20 s, and achieving less than 3% optimality gap for large instances with more than 250,000 decision variables in less than 10 min on average. We also find that in addition to problem size, the polyhedral properties of a UFLP instance may significantly impact the efficiency of quantum annealing based algorithm.