A Fast Swap-based Local Search Procedure for Location Problems

We present a new implementation of a widely used swap-based local search procedure for the p-median problem, proposed in 1968 by Teitz and Bart. Our method produces the same output as the best alternatives described in the literature and, even though its worst-case complexity is similar, it can be signicantly faster in practice: speedups of up to three orders of magnitude were observed. We also show that our method can be easily adapted to handle the facility location problem and to implement related procedures, such as path-relinking and tabu search.