4
$\begingroup$

I have been trying to practice recurrence relations that can be solved by the master theorem and came across this.

Now the $4^{\textrm{th}}$ problem in that file is :

$$T(n) = 2^n T\left(\frac{n}{2}\right) + n^n.$$

The solution says it can't be solved using Master theorem, which is true. But I want to know how to solve it.

I tried thinking of some substitution but I'm not able to get anywhere since the recurrence contains terms $2^n$ and $n^n$ both. And if I expand the tree, it becomes really messy. Any solutions?

$\endgroup$
2
$\begingroup$

Introducing $U(n)=T(n)/4^n$, one sees that $U(n/2)=T(n/2)/2^n$ hence the recursion over $(T(n))_n$ yields $U(n)=U(n/2)+(n/4)^n$.

Roughly speaking, this means that $U(n)=(n/4)^n+(n/8)^{n/2}+(n/16)^{n/4}+\ldots$ More rigorously, the $\ldots$ hides at most $\log_2n$ terms, each less than $(n/8)^{n/2}$, hence $$ (n/4)^n\leqslant U(n)\leqslant(n/4)^n+\log_2(n)\cdot(n/8)^{n/2}, $$ which implies $$ n^n\leqslant T(n)\leqslant n^n+\log_2(n)\cdot(2n)^{n/2}. $$ Since $\log_2(n)\cdot(2n)^{n/2}\ll n^n$, this yields $\lim\limits_{n\to\infty}T(n)/n^n=1$, hence $T(n)=\Theta(n^n)$.

$\endgroup$
1
$\begingroup$

Let $\begin{cases}k=\log_2n\\U(k)=T(n)\end{cases}$ ,

Then $U(k)=2^{2^k}U(k-1)+(2^k)^{2^k}$

$U(k)=2^{2^k}U(k-1)+2^{k\times2^k}$

$U(k+1)=2^{2^{k+1}}U(k)+2^{(k+1)2^{k+1}}$

$\therefore U_c(k+1)=2^{2^{k+1}}U_c(k)$

$U_c(k)=\Theta(k)\prod\limits_k2^{2^{k+1}}$ , where $\Theta(k)$ is an arbitrary periodic function with unit period

According to http://en.wikipedia.org/wiki/Indefinite_product#Rules ,

$U_c(k)=\Theta(k)2^{\sum\limits_k2^{k+1}}$ , where $\Theta(k)$ is an arbitrary periodic function with unit period

According to http://en.wikipedia.org/wiki/Indefinite_sum#Antidifferences_of_exponential_functions ,

$U_c(k)=\Theta(k)2^{2^{k+1}}$ , where $\Theta(k)$ is an arbitrary periodic function with unit period

Let $U(k)=2^{2^{k+1}}V(k)$ ,

Then $2^{2^{k+2}}V(k+1)=2^{2^{k+1}}2^{2^{k+1}}V(k)+2^{(k+1)2^{k+1}}$

$2^{2^{k+2}}V(k+1)=2^{2^{k+2}}V(k)+2^{(k+1)2^{k+1}}$

$V(k+1)=V(k)+2^{(k-1)2^{k+1}}$

$V(k)=\Theta(k)+\sum\limits_k2^{(k-1)2^{k+1}}$ , where $\Theta(k)$ is an arbitrary periodic function with unit period

$U(k)=\Theta(k)2^{2^{k+1}}+2^{2^{k+1}}\sum\limits_k2^{(k-1)2^{k+1}}$ , where $\Theta(k)$ is an arbitrary periodic function with unit period

$T(n)=\Theta(\log_2n)4^n+4^n\left(\sum\limits_k2^{(k-1)2^{k+1}}\right)_{k\to\log_2n}$ , where $\Theta(n)$ is an arbitrary periodic function with unit period

$\endgroup$
-1
$\begingroup$

Hint: if $k = T(n)/n^n$ then $$T(2n) = k 2^{2n}n^n +(2n)^{2n} = \left(k/n^n+1\right) (2n)^{2n}$$

$\endgroup$
  • $\begingroup$ And? $ $ $ $ $ $ $\endgroup$ – Did Oct 2 '12 at 10:39
  • $\begingroup$ and $T(n) \in \Theta(n^n)$ as for large $n$ you have $k/n^n+1$ getting much closer to $1$ than $k$ $\endgroup$ – Henry Oct 2 '12 at 10:48
  • $\begingroup$ This shows that T(2n)/(2n)^(2n) converges to 1 as soon as T(n)/n^(2n) converges to 0. Hence one needs to show that T(n)/n^(2n) does converge to 0. $\endgroup$ – Did Oct 2 '12 at 10:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.