On The Travelling Salesman Problem with Neighborhoods in a Polygonal World

This page contains dataset instances for The Travelling Salesman Problem with Neighborhoods in a polygonal world that were used in our CLAWAR 2022 paper. As there is no dataset for the mentioned problem, two sets of instances were created. The first set A consists of six instances, with the same polygonal map of 59 obstacles and the varying number n of neighborhoods, n ∈ {12, 24, 33, 59, 99, 224}. The neighborhoods were generated to almost cover the whole map, so they have varying numbers of vertices and various shapes.

The second set B is based on the same map with 59 obstacles. We generated 150 random points in the map from which we computed visibility regions and cut them by regular decagons centered in these points to simulate a limited visibility range. 150 polygons of a similar size with around ten vertices were generated. We gradually removed polygons from this set of polygons to get instances with n = {10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 130, 150} neighborhoods.

Set A - Covering

Set B - Limited visibility

Download the dataset here.