0
$\begingroup$

Which of the following is false?

A) $100n \log{n}=\left(\frac{n \log{n}}{100}\right)$

B) $\sqrt{\log{n}}=O(\log \log{n})$

C) If $0

D) $2^n \neq O(nk)$

What is the correct answer here? I'm sure $C$ and $D$ are right but not able to judge between $A$ and $B$. This is a GATE Exam question.

  • 0
    There's no $O$ in **A.**. As for **D.**, isn't it $O(n^k)$? – 2017-01-13
  • 0
    ohh sorry ya it is there . – 2017-01-13
  • 0
    where i can ask question about program output ? – 2017-01-13
  • 0
    Which program output? I don't understand. – 2017-01-13
  • 0
    http://math.stackexchange.com/questions/2095988/how-these-two-function-notations-are-same – 2017-01-13

2 Answers 2

1

For B and similar problems with mulitple logs, exponentials, and powers, a trick I've found helpful is taking an extra log or two of both sides.

Instead of asking whether $\sqrt{\log n} = O(\log \log n)$, investigate whether $\frac{1}{2} \log \log n = O( \log \log \log n )$, which is of course false since the right side has more logs and thus will grow asymptotically slower. Thus, the original claim is also false.

  • 0
    its good method . bt today i posted similar question . where 1.0000001 ^n and n^7/4 now tell which is bigger ? – 2017-01-13
  • 0
    http://math.stackexchange.com/posts/comments/4307906?noredirect=1 – 2017-01-13
  • 0
    Applying this method to that question would frame it as "Which is eventually bigger, $n \log 1.0000001$ or $\frac{7}{4} \log n$?" after you apply some properties of logarithms. Is that clearer? – 2017-01-13
  • 0
    little help need . so after implementing ur method second term is bigger ? is it right? – 2017-01-13
  • 0
    $\log n$ grows slower than any power function $n^k$, so the first term is bigger. $\log 1.0000001$ is very small and makes the first term smaller for small $n$, but by the time $n$ is $10^{20}$, the functions reveal their true colors. – 2017-01-13
0

A: You even can compute the ratio of the functions, not only bound it.

B: Is $\;\dfrac{\log n}{\log\log n}\;$ bounded? (Hint: $\log u=_\infty o(u)$).

D: Is $\;\dfrac{2^n}{n^k}\;$ bounded?

  • 0
    I think a is right bcz we can write it like n log n = O ( n log n / 10000) and generally we neglect the constant . – 2017-01-13
  • 0
    but wht about b ? its not clear in my thinking ...cn ny one tell better reason else i need to learn it . My calculator also give nearly equal answer in both ones . – 2017-01-13
  • 0
    For A: yes, we can remove the constant factors, by definition. For B: do you know what $\;\log u=o(u)\;$ mean? – 2017-01-13
  • 0
    ya it means log u < o(u) , not equal it . – 2017-01-13
  • 0
    Not exactly: it means the ratio s $\dfrac{\ln u}u$ tends to $0$ as $u$ tends to $+\infty$. Apply this with $u=\log n$. – 2017-01-13