-4

How would you go about proving $n! < n^n$ using a mathematical induction proof?

I understand how to solve inductive proofs with = but I'm getting a bit lost in this example.

Any help is much appreciated.

jknam
  • 5
  • You probably mean either $\leq$ or you want to prove the statement for $n\geq 2$... – 5xum Oct 27 '15 at 20:19
  • Does one really need induction to see that $n! = 1 \cdot 2 \cdots n \le n \cdot n \cdots n = n^n$ ? – lhf Oct 29 '15 at 17:37
  • @lhf it's not a question of do you really need t to see that it's true, it's a question of proving it. – jknam Oct 29 '15 at 17:48

2 Answers2

2

First of all, we have $$ 2! < 2^2 $$ which covers the base case. Now, for the induction itself, assume we know that $k!<k^k$ for some $k$. We want to compare $(k+1)!$ and $(k+1)^{k+1}$. Note that $(k+1)! = k!\cdot (k+1)$, which lets us use the induction hypothesis (I.H.): $$ (k+1)! = k!\cdot (k+1) \overset{\text{I.H.}}{<} k^k\cdot(k+1) \overset{k<k+1}{<} (k+1)^k\cdot (k+1) = (k+1)^{k+1} $$ which is what we want.

Arthur
  • 199,419
  • So in the part where you said $k^k * (k+1) < (k+1)^k*k+1$ why did you not substitute k+1 in the exponent of $(k+1)^k$?That is the part I am confused about – jknam Oct 27 '15 at 20:37
  • @jknam When I do that transition, I'm really doing the transition $k^k < (k+1)^k$, just with both sides multiplied by $k+1$. While it is true that $k^k < (k+1)^{k+1}$ also, when put into the context of the whole inequality, that would give us $(k+1)! < (k+1)^{k+1}\cdot (k+1) = (k+1)^{k+2}$ which is not what we want. – Arthur Oct 27 '15 at 20:52
  • Right so how did you determine when you started solving this problem that you wouldn't want to substitute completely? – jknam Oct 27 '15 at 21:46
  • @jknam A general rule is to substitute as little as possible so that we actually end up with a stronger statement ($(k+1)! < (k+1)^{k+1}$ is stronger than $(k+1)! < (k+1)^{k+2}$) but still be able to get a nice and self-contained result (too many terms to make the inequality closer to equality, and it just becomes uninteresting). In this specific case, I already had $(k+1)^1$, and I wanted to end up with $(k+1)^{k+1}$, which means that I needed whatever else there was to become exactly $(k+1)^k$. Otherwise it wouldn't've matched up to what I wanted in the end. – Arthur Oct 27 '15 at 21:54
  • I see. So as a rule of thumb, substitute to match the IH? – jknam Oct 27 '15 at 21:55
  • @jknam Definitely (or rather, to match the expected result of the induciton step; the IH is $k! < k^k$, while we want to end up at $(k+1)! < (k+1)^{k+1}$). See where you want to end up, see what you need to end up where you want, and try to make that happen. – Arthur Oct 27 '15 at 21:56
0

base case has to be 2 which is evident $$2!=2<2^2=4$$ assume $n!<n^n$, now we have $$(n+1)!=(n+1)n!<(n+1)n^n=n^{n+1}+n^n<n^{n+1}+n^{n+1}=2n^{n+1}<(n+1)^{n+1}$$

Zelos Malum
  • 6,570
  • So I'm not following where you say $n^(n+1) + n^n < n^(n+1) + n^(n+1)$, where do the n^(n+1) 's on the right side come from? Thanks again for answering my question though, I really appreciate it. – jknam Oct 27 '15 at 20:25