A process for solving constrained optimization problems
Based on various optimization problems I have solved on various classical, QUBO-based solvers, annealers, and gate-quantum computers, it is important to use a step-by-step process where there is a good understanding of the problem classically before moving on to developing the QUBO. When developing the QUBO formulation, it is important to get the constraint strengths right, which may require some trial and error or calculations to get the optimal multipliers. If a QUBO formulation that results in a single matrix is not developed correctly, none of the QUBO solvers, annealers, or gate-quantum computing methods will be able to give the best result. So, it is important to get this formulation and resulting matrix correct. Then, before the matrix can be used by various QUBO solvers, it is important to understand the format the solver requires. For example, for the D-Wave annealers, the QUBO needs to be converted into a BQM before...