You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently random_configuration_model uses rejection sampling which is very slow for tight configurations. Probably allowing for approximate sampling with some edge rewiring strategy could be allowed (iGraph implements this).
The text was updated successfully, but these errors were encountered:
I already noticed that the algorithm was not the true random_configuration_model as it does not allows the creation of multigraphs. If I remember correctly, even with the rejection sampling used, I think that the current generation is not exactly uniform (to be checked), so maybe it makes sense to do that...
I assumed it was on purpose as SimpleGraph does not support multigraphs
current generation is not exactly uniform
Ah - indeed - I have just checked. It does some "naive" collision resolution breaking uniformity. So indeed we are neither uniform nor "fast near uniform" now.
Follow up to https://stackoverflow.com/questions/72732571/random-configuration-modeln-e-takes-to-long-on-lightgraphs-jl.
Currently
random_configuration_model
uses rejection sampling which is very slow for tight configurations. Probably allowing for approximate sampling with some edge rewiring strategy could be allowed (iGraph implements this).The text was updated successfully, but these errors were encountered: