Quadratic Assignment

QAP is a long standing optimization problem which is strongly NP-Hard. Despite of extensive research in this field, there are still some problems in the QAPLIB benchmarking datasets that have not yet been optimally solved. My approach is to apply the IP solver framework with the Sinkhorn approach as the lower bounding scheme, and Branch and Bound (B&B) as the gap closure scheme. Many real life applications are modeled as QAPs. I primarily focus on the shape matching problem in the image processing domain and the facility location in the Operations Research (OR) domain.