Approximating the maximum quadratic assignment problem

In the maximum quadratic assignment problem, two nonnegative symmetric matrices are given and the objective is to compute a permutation. The problem is NP-hard and an indication to the hardness of approximating it, is that the best known approximation factors for maximum clustering with given sizes are 1/c when all sizes are equal to a constant c. An approximation algorithm with a constant performance guarantee under the assumption that the weights in B satisfy the triangle inequality.