关于Chimera图上定义的QUBO实例的注记
A note on QUBO instances defined on Chimera graphs
摘要:
McGeoch和Wang(2013)最近使用439量子位D-Wave Two量子计算系统在比使用IBM BLOG CPLEX混合整数二次规划(MIQP)求解器少得多的时间内获得了一些二次无约束布尔优化(QUBO)问题实例的最优或接近最优解。McGeoch和Wang研究的问题是在Chimera图的子图上定义的-最多有439个节点。我们观察到,在将QUBO问题标准地重新表述为混合整数线性规划(MILP)之后,McGeoch和Wang使用的特定实例可以用CPLEX MILP求解器在比McGeoch和Wang报告的CPLEX MIQP求解器的时间少得多的时间内求解到最优。然而,解决的时间仍然比McGeoch-Wang测试中的D-Wave计算机所花费的时间更长。
英文摘要:
McGeoch and Wang (2013) recently obtained optimal or near-optimal solutions to some quadratic unconstrained boolean optimization (QUBO) problem instances using a 439 qubit D-Wave Two quantum computing system in much less time than with the IBM ILOG CPLEX mixed-integer quadratic programming (MIQP) solver. The problems studied by McGeoch and Wang are defined on subgraphs -- with up to 439 nodes -- of Chimera graphs. We observe that after a standard reformulation of the QUBO problem as a mixed-integer linear program (MILP), the specific instances used by McGeoch and Wang can be solved to optimality with the CPLEX MILP solver in much less time than the time reported in McGeoch and Wang for the CPLEX MIQP solver. However, the solution time is still more than the time taken by the D-Wave computer in the McGeoch-Wang tests.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。