New Heuristic Algorithms for Solving Single-Vehicle and Multi-Vehicle Generalized Traveling Salesman Problems (GTSP)

摘要

Among numerous NP-hard problems, the Traveling Salesman Problem (TSP) has been one of the most explored, yet unknown one. Even a minor modification changes the problema€?s status, calling for a different solution. The Generalized Traveling Salesman Problem (GTSP)expands the TSP to a much more complicated form, replacing single nodes with a group or cluster of nodes, where the objective is to find a minimum-length tour containing exactly one node from each cluster. In this paper, a new heuristic method is presented for solving singlevehicle single-depot GTSP with the ability of controlling the search strategy from conservative to greedy and vice versa. A variant algorithm is then developed to accommodate the multi-vehicle single-depot condition, which is modified afterwards to accommodate the multi-vehicle multi-depot GTSP.