I'm interesting in generating directed acyclic graphs (see here, for example). As part of this search, I'm curious if there are any efficient algorithms for determining a canonization of a directed acyclic graph. Or, are there any tricks that allow the canonization of a dag to be computed more efficiently than a general graph?
Determining the canonical form in general is GI-complete, but there are many specialized graphs for which it can be calculated in polynomial time. I didn't have any luck determining whether acyclic digraphs were among those specialized types, however.