“…In principle, it would be possible to use the exact methods developed in some prior studies in parallel: run the algorithm for a given number of patterning steps and then verify for which number of patterning steps zero coloring conflicts emerge. However, most of these methods either employ heuristics to accelerate finding a coloring solution at a large scale (and hence no longer guaranteeing optimality) [37,43], propose formulations that do not work for any number of patterning steps [5,26], or exploit additional structures and/or placement options [12,26,38]. In our case, given that the objective is to formally find the minimum number of patterning steps required for (large scale) layouts, we do not build on the methods developed in prior research.…”