Deterministic, stochastic and local search algorithms for real-world ATSP instances

This paper considers six asymmetric travelling salesman problem instances generated from the real data. To solve these instances, this paper develops multiple deterministic, stochastic and local search algorithms which construct their solutions based on the geographical information, such as the adjacent cities, the border cities, the boundary of the country, and the directions on map. The neighbourhood structures of the proposed local search algorithms are very efficient and contain only small numbers of solutions. All proposed algorithms are simple and very fast.