I wonder if the enumeration approach from this paper : https://pubs.acs.org/doi/10.1021/ci025526c# https://pubs.acs.org/doi/10.1021/ci025526c# which (I'm awar

I wonder if the enumeration approach from this paper : https://pubs.acs.org/doi/10.1021/ci025526c# https://pubs.acs.org/doi/10.1021/ci025526c# which (I'm aware it's paywalled) essentially enumerates using the inner duals of the tilling and then reconstructs from that inner dual. Seems like if you could just run through the legal tilings, rather than have to reject ones where the random choice of next tile was not legal, it would be faster