📝 SummarXiv

Abstract

Removing overlaps is a central task in domains such as scheduling, visibility, and map labelling. This task can be modelled using graphs, where overlap removals correspond to enforcing a certain sparsity constraint on the graph structure. We continue the study of the problem Geometric Graph Edit Distance, where the aim is to minimise the total cost of editing a geometric intersection graph to obtain a graph contained in a specific graph class. For us, the edit operation is the movement of objects, and the cost is the movement distance. We present an algorithm for rendering the intersection graph of a set of unit circular arcs (i)~edgeless, (ii)~acyclic, and (iii)~$k$-clique-free in $O(n\log n)$ time, where $n$ is the number of arcs. We also show that the problem remains strongly NP-hard on unweighted interval graphs, solving an open problem of [Honorato-Droguett et al., WADS 2025]. We complement this result by showing that the problem is strongly NP-hard on tuples of $d$-balls and $d$-cubes, for any $d\ge 2$. Finally, we present an XP algorithm (parameterised by the number of maximal cliques) for rendering the intersection graph of a set of weighted unit intervals edgeless.