Constraint Mixers for Optimization Problems on Circuit Model Quantum Computing

In quantum optimization, the search space with standard X-mixers for Quantum Approximate Optimization Algorithm (QAOA) and Trotterized Adiabatic Evolution often includes infeasible subspaces. This can lead to infeasible final solutions with circuit model quantum algorithms. Utilising constraint mixers for some or all constraints in the problem allows the search within the feasible subspace and hence offers better convergence to good solutions. In this talk, we will present several constraint mixers designed for common constraints found in various optimization problems and share results on different optimization problems along with the comparison of constraint mixer approaches over the standard approach in improving solution quality on circuit model quantum computing.