Reconfiguration in Curve Arrangements to Reduce Self-Intersections and Popular Faces

Abstract

We study reconfiguration in curve arrangements, where a subset of the crossings are marked as switches which have three possible states, and the goal is to set the switches such that the resulting curve arrangement has few self-intersections, or few faces that are incident to the same curve multiple times (a.k.a. popular faces). Our results are that these problems are NP-hard, but FPT in the number of switches. Minimizing self-intersections is also FPT in the number of non-switchable crossings; for minimizing popular faces this problem remains open. Our results can be applied to generating curved nonograms, a type of logic puzzle that has received some attention lately. Specifically, our results make it possible to efficiently convert expert puzzles into advanced puzzles (or determine that this is impossible).