Sufficient optimal conditions for unconstrained quadratic binary problems

In this article, we present several sufficient optimal conditions for unconstrained quadratic binary problems, which can be applied in algorithms combining with SDP relaxations in branch-and-bound approaches for the primal problem. These optimal conditions can work for many situations when the Lagrangian duality gap is not zero.