What are the general attributes of an algorithm that cause a polylogarithmic term (e.g. O(log log N)
) to be introduced into its time or space complexity?
I'm asking about general structural or behavioral properties, e.g. O(N^3)
often corresponds to a triply-nested loop over all elements in a collection, or O(log N)
often indicates that the algorithm could be modeled as a process of walking a path from the root to a leaf in a balanced tree.