Is it possible for QBSolve to return an infeasible solution to the problem, even though matrix Q is guaranteed to be correct?

0

I'm modeling a QUBO problem, using D-Wave's QBSolve to return a solution to an original optimization problem P, and I'm aware that matrix Q matches problem P perfectly. I already know that QBSolve doesn't guarantee optimality, but it can return viable solutions. However, I'm coming across infeasible solutions. Is it possible for QBSolve to return an infeasible solution to the problem, even though matrix Q is guaranteed to be correct? If yes, why does this happen? And what are the current alternatives for working with huge matrices that only return viable solutions to the problem? I'm waiting for an answer to my question.

asked 2 years ago247 views
1 Answer
1
Accepted Answer

Hello! QBSolv is now deprecated and it is recommended by D-Wave to use dwave-hybrid.

QBSolv makes uses of the tabu algorithm when running on classical solvers. More information can be found on the docs. Infeasible solutions have not been reported before but I would recommend looking into dwave-hybrid. Let us know if this answers your question!

AWS
Abe C
answered 2 years ago
  • Thanks for your response. In light of current dwave-hybrid approaches, I have used the D-wave Simulator for simulated annealing (neal - https://docs.ocean.dwavesys.com/projects/neal/en/latest/reference/sampler.html). Like QBSolv, it is also returning infeasible solutions. In the case of D-wave simulators, are they suitable for large problems, and can they produce unfeasible solutions? I'm sure I'm working with the correct QUBO array for the problem.

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions