Do use Solvers for QUBO problems always return viable solutions to the problem?
0
Hi, I'm currently working on problems that I turn into a Quadratic Unconstrained Binary Optimization (QUBO) problem and solve them with the D-wave plugin on Amazon Braket to return viable solutions.
Once I have constructed the matrix Q correctly given the original linear problem, is there a guarantee that if I use the quantum computer with Quantum Annealing to generate solutions to the original problem, the result that the simulator returns will always be a viable solution? Or is it possible for the simulator to return feasible and infeasible solutions to the modeled problem?
The question arose when I submitted the Q matrix to the D-Wave simulator on Amazon Braket. The returned solution is not feasible for my problem (i.e., one of the constraints is being violated). Or can this happen?