[This page is auto-generated; please prefer to read the article’s PDF Form.]
For a function h ∈ℋ, let us denote by Si the set of keys from S which are mapped to i ∈ ℤm by h. Say si = |Si|. So, the number of colliding-pairs in slot i is . Hence,
So the following inequality from theorem 3 can be rewritten:
Corollary 4. Given 𝜖-Universal class ℋ and any set of n keys, if si is the number of keys mapped to slot i by a function h ∈ℋ, at least half of the functions in class ℋ must have si such that:
|
This upper-bound on the sum of si2 finds its use for optimizing space in the FKS Method of Perfect Hashing [1].
■
[1] M. L. Fredman, J. Komlós, E. Szemerédi. Storing a Sparse Table with O(1) Worst Case Access Time. J. ACM, Vol 31 (3) (1984), 538–544.
[2] Z. J. Czech, G. Havas, B. S. Majewski. Fundamental Study: Perfect Hashing. Theoretical Computer Science, Vol 182 (1–2) (1997), 1–143.
[3] T. H. Cormen, C. E. Leiserson, R. L. Rivest, C. Stein. Introduction to Algorithms, Third Edition. The MIT Press, 2009.
Copyright © 2021 Nitin Verma. All rights reserved.