Abstract
This paper is concerned with a new approach for solving quadratic assignment problems (QAP). We first reformulate QAP as a concave quadratic programming problem and apply an outer approximation algorithm. In addition, an improvement routine is incorporated in the final stage of the algorithm. For large size problems, we examined a Simulated Annealing as the final stage of algorithm. Computational experiments on two series of standard data demonstrate that this algorithm can yield favorable results within reasonable computational effort.
