algorithmtime-complexitybig-ohashtableclrs

The interpretation of expected time bound for searches in a hash table


As CLRS book,page 260 stated,

enter image description here

I wouldn't have any problem if the author says the bound is eventually
enter image description here or even
enter image description here
What kind of theories shall we apply to simplify the original result, i.e, cancelling the factor 1/n of a. Since n is one of inputs of the function, is it necessary to cancel it by treating it as a constant? What i've missed? is anyone got the same confusion T_T?


Solution

  • alpha/n is asimptotically smaller (has lower order) being compared with alpha, so it might be ignored. When n (hashtable size, AFAIU) becomes larger, 1/n value tends to zero.
    Note - wiki table does not contain 1/n function because it is evaluated as having zero impact

    Alike situation - if time is Theta(n^2 + 100*n + 10000), dominating summand is quadratic, and

    Theta(n^2 + 100*n + 10000)` = Theta(n^2)