Hacker News new | past | comments | ask | show | jobs | submit login

Is there some result in the theory of random graphs that says that you can't have a graph of a certain size without developing hubs? It seems to happen in most systems.



There are a number of results about a giant connected component emerging as a Erdos-Renyi random graph increases in size.

A hub however will simply arise as the result of variance.


You say "a hub" but I wonder whether it would be many? How close would such an effect be to hubs we see in scale free networks? In those, preferential attachment seems to induce hubs. Airline hubs are a good example - there's more value in connecting to a well-connected hub.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: