Solving Large Quadratic Assignment Problems

Quadratic Assignment problems are in practice among the most diicult to solve in the class of NP-complete problems. The only successful approach hitherto has been Branch-and-Bound-based algorithms, but such algorithms are crucially dependent on good bound functions to limit the size of the space searched. Much work has been done to identify such functions for the QAP, but with limited success. Parallel processing has also been used in order to increase the size of problems solvable to optimality. The systems used have, however, often been systems with relatively few, but very powerful vector processors, and have hence not been ideally suited for computations essentially involving non-vectorizable computations on integers. In this paper we investigate the combination of one of the best bound functions for a Branch-and-Bound algorithm (the Gilmore-Lawler bound) and various testing, variable binding and re-calculation of bounds between branchings when used in a parallel Branch-and-Bound algorithm. The algorithm has been implemented on a 16-processor MEIKO Computing Surface with Intel i860 processors. Computational results from the solution of a number of large QAPs, including the classical Nugent 20 are reported.