Ant Colony System Based Approaches to the Air-Express Courier's Routing Problem

Abstract

This paper proposes two ant colony system (ACS) based approaches, called stepwise-ACS (SACS) and cheapest-insertion-stepwise-ACS (CISACS), to solve the air-express courier's routing problem. The courier visits N predetermined delivery points and M pickup requests during the en route delivery. The SACS performs an ACS to obtain the initial tour for the N delivery points and then to find the shortest Hamiltonian path, in a stepwise manner, when the M requests arrive. The CISACS, following the SACS, incorporates the cheapest insertion into the SACS as new requests become known. Experiments with various pickup emergence patterns that characterize the real-world circumstances are tested for the proposed algorithms. The computational results, in terms of total traveled distance, are compared with the baseline results by a cheapest insertion (CI) heuristic. The results show that both SACS and CISACS perform better than the CI heuristic and that CISACS yields the lowest traveled distance.