Cover Data Structures and Algorithms with Object-Oriented Design Patterns in Java
next up previous contents index

About Logarithms

In this section we determine the asymptotic behavior of logarithms. Interestingly, despite the fact that tex2html_wrap_inline58508 diverges as n gets large, tex2html_wrap_inline58512 for all integers tex2html_wrap_inline57996. Hence, tex2html_wrap_inline58516. Furthermore, as the following theorem will show, tex2html_wrap_inline58508 raised to any integer power tex2html_wrap_inline58520 is still O(n).

Theorem  For every integer tex2html_wrap_inline58520, tex2html_wrap_inline58526.

extbfProof This result follows immediately from Theorem gif and the observation that for all integers tex2html_wrap_inline58520,

  equation1613

This observation can be proved by induction as follows:

Base Case Consider the limit

displaymath58504

for the case k=1. Using L'Hôpital's rulegif  we see that

eqnarray1635

Inductive Hypothesis Assume that Equation gif holds for tex2html_wrap_inline58550. Consider the case k=m+1. Using L'Hôpital's rule  we see that

eqnarray1648

Therefore, by induction on m, Equation gif holds for all integers tex2html_wrap_inline58520.

For example, using this property of logarithms together with the rule for determining the asymptotic behavior of the product of two functions (Theorem gif), we can determine that since tex2html_wrap_inline58516, then tex2html_wrap_inline58560.


next up previous contents index

Bruno Copyright © 1998 by Bruno R. Preiss, P.Eng. All rights reserved.