To solve problems of quadratic binary programming, we suggest a hybrid approach based on the cooperation of a new version of DCA (Difference of Convex functions Algorithm), named the DCA-Like, a Genetic Algorithm and Migrating Bird Optimization algorithm. The component algorithms start in a parallel way by adapting the Master-Slave model. The best-found solution is distributed to all algorithms by using the Message Passing Interface (MPI) library. At each cycle, the obtained solution serves as a starting point for the next cycle’s component algorithms. To evaluate the performance of our approach, we test on a set of benchmarks of the quadratic assignment problem. The numerical results clearly show the effectiveness of the cooperative approach.