My understanding is the requirement an optimal route be a simple polygon is based on the use of Euclidean distance.I had the idea to test whether the tour is a simple polygon, and do improvements before the recreate steps if not.
When Euclidean distance is rounded to the nearest integer, could the rounding operations be arranged in such a way the optimal route has crossing? This seems plausible to me.
Statistics: Posted by ejolson — Tue Sep 02, 2025 2:58 am