On parallel local search for permutations

We investigate some ways in which massively parallel computing devices can be exploited in local search algorithms. We show that the substantial speed-ups that can be gained from parallel neighbourhood evaluation enables an efficient best improvement local search, and that this in turn enables further speed-ups through selection and parallel application of a set of independent, improving moves. Our experiments demonstrate a total speed-up of up to several hundred times compared to a classical, sequential best improvement search. We also demonstrate how an exchange of good partial solutions between the incumbent and best found solutions improves the efficiency of the Iterated Local Search