Διαχωρισμένη τοπική αναζήτηση για το πρόβλημα ομαδικού προσανατολισμού με χρονικά παράθυρα
Partitioned local search for the team orienteering problem with time windows

View/ Open
Abstract
The current work extensively studies the Team Orienteering Problem with Time Windows, which is an extension of the Orienteering Problem. The Orienteering Problem belongs to NP-hard problems, which makes it impossible to solve in polynomial time for large input data. For this reason, the use of heuristic and approximate algorithms becomes necessary to find satisfactory solutions in a short period of time. Several algorithms have already been implemented for the Orienteering Problem, one of which is the Iterated Local Search (ILS) algorithm. The purpose of this paper is to reduce the execution time of ILS by partitioning the problem graph in a satisfactory way, applying a partitioned Local Search to the individual sub-graphs, and dealing with the problems arising from this partitioning. The experimental results show the reduction of the execution time of the algorithm especially in examples with large input data, but with the impact of the reduction of the score of the solutions. Nevertheless, steps have been taken to maintain the solutions at a satisfactory level, while there is also room for further improvement.