1

For which $n\geq0$ is $3^n>n^3$ valid? Consider the four first cases: $$3^0=1>0$$ $$3^1=3>1^3$$ $$3^2=9>2^3$$ $$3^3=27\ngtr3^3$$

So let us consider the base case $k=4$:

$$3^4=81 >4^3,$$

which holds. Then we need to show $3^{k+1}>(k+1)^3 $

$3k^3>(k+1)^3\iff \left(1+\frac1k\right)^3<3$

For $k=3,\left(1+\frac1k\right)^3=\frac{64}{27}<3$

and $\left(1+\frac1{k+1}\right)^3<\left(1+\frac1k\right)^3$

$\implies \left(1+\frac1k\right)^3<3$ for $k\ge3$

Hence for the values $n=1,2$ and for $n>3$, $3^n>n^3$ holds. For $n=3$ we have strict equality, that is $3^n=n^3$. \

But there is a problem,

$3k^3>(k+1)^3\iff \left(1+\frac1k\right)^3<3$

is probably incorrect, but it gives the right result!

What is the correct step of this induction?

Luthier415Hz
  • 2,739
  • 6
  • 22
  • 1
    I'm not sure if you want to use $3^{k+1} = 3k^3$, which is wrong. I think what you might mean is $3^{k+1} = 3 \cdot 3^k$? – Targon Jul 11 '23 at 09:37
  • 4
    @Luthier415Hz First of all, where's your base case? Where's your induction hypothesis? Only once those are in place can you introduce the inductive step. It seems like a very badly-written question; have you given us everything you have on it? For what range of $k$ is this result supposed to hold? You determined $k\ge 3$, but in fact $k=0$ works as well – H. sapiens rex Jul 11 '23 at 09:37
  • @H.sapiensrex updated OP – Luthier415Hz Jul 11 '23 at 09:42
  • $3^4>4^3$ and for $k\geq4$, if $3^k>k^3$ then $$3^{k+1}=3\cdot3^k>3\cdot k^3>(k+1)^3.$$ – Servaes Jul 11 '23 at 22:00

2 Answers2

4

$\rm {Task :}$

Let $n$ be a positive integer . Prove that, $3^n>n^3$ holds $\forall n≥4$ by induction .


For $n=4$, the statement is obviously correct .

Suppose that, the statement is correct for $n=k≥5$. Then we can write :

$$3^k-k^3=a,\thinspace a>0$$

$\rm{Inductive \thinspace\thinspace step\thinspace\thinspace .}$

For $n=k+1$ we have :

$$ \begin{align}&\thinspace\thinspace\thinspace3^{k+1}-(k+1)^3\\ =&\thinspace\thinspace\thinspace 3(a+k^3)-(k+1)^3\\ >&\thinspace\thinspace\thinspace 3k^3-(k+1)^3\\ ≥&\thinspace\thinspace\thinspace 3k^3-\left(k+\frac k5\right)^3\\ =&\thinspace\thinspace\thinspace \frac{159}{125}k^3\thinspace >\thinspace 0\end{align} $$

which completes the proof .


$\rm {Alternative \thinspace \thinspace methods \thinspace\thinspace .}$

You can also observe that, since

$$ \begin{align}&\sqrt [3]{3}-1>\frac 15\\ \iff &3>\left(1+\frac 15\right)^3\end{align} $$

Then, we have :

$$ \begin{align}&3k^3>(k+1)^3\\ \iff &k\sqrt [3]3>k+1\\ \iff &\underbrace k_{\color{#0a0}{≥}\color{#c00}{\thinspace 5}}\thinspace\thinspace\underbrace{\left(\sqrt [3]3-1\right)}_{\color{#0a0}{>}\color{#c00}{\thinspace\frac 15}}>1\thinspace\thinspace .\end{align} $$


Note that, we can also complete the proof using the method of Polynomial long division or Synthetic division :

$$ \begin{align}&\thinspace\thinspace\thinspace 3k^3-(k+1)^3\\ =&\thinspace\thinspace\thinspace 2k^3-3k^2-3k-1\\ =&{\thinspace\thinspace\thinspace\underbrace {(k\!-\!5)}_{\color{#0a0}{≥}\color{#c00}{\thinspace 0}}\thinspace\underbrace {(2k^2\!+\!7k\!+\!32)}_{\color{#0a0}{>}\color{#c00}{\thinspace 0}}\!+\!159}\\ ≥&\thinspace\thinspace\thinspace 159\thinspace>\thinspace 0\thinspace\thinspace .\end{align} $$

lone student
  • 14,709
3

There are three parts of the proof by induction:

  1. Base case: the first number considered which fulfils the property/equation/inequality to be proved. Here the base case is $n=4$, and the inequality to be proved is $3^n>n^3$, which is fulfilled by $n=4$ (because $81>64$). Note that we use the variable $n$ here, the same as in the original inequality. We do not need to use $k$ yet in this step.

  2. Induction step: assuming the case $n=k$ is true, for any $k\ge 4$, we have to prove that the case $n=k+1$ is true.

First we assume the case $n=k$ is true, that is, we assume $$3^k>k^3.$$ From that assumption, we try to prove that $3^{k+1}>(k+1)^3$. Let's start by multiplying the assumed inequality by $3$: $$3^{k+1}>3k^3$$ Now consider that for every $k\ge 4$, $$\left(\frac{k+1}{k}\right)^3<3$$ which is equivalent to $$3k^3>(k+1)^3.$$ From the two inequalities obtained, we conclude $$3^{k+1}>(k+1)^3,$$ proving that the case $n=k+1$ holds.

  1. The conclusion: As we see that the case $n=4$ holds, and for every $k\ge4$, the inequality $3^k>k^3$ implies $3^{k+1}>(k+1)^3$, we conclude that the inequality $3^n>n^3$ holds for every natural number $n\ge4$.
user_194421
  • 2,291
  • 3
    You didn't prove $3k^3>(k+1)^3$ Why $\left(\frac{k+1}{k}\right)^3<3$ ? – nonuser Jul 11 '23 at 13:18
  • 2
    @nonuser: There are probably easier ways (e.g. see lone student's answer), but here's one way (I was curious, so I did this on scratch paper) -- $3k^3 > (k+1)^3$ is equivalent to $k^3 + k^3 > 3k^2 + 3k + 1,$ so it suffices to prove this last inequality, which follows from $k^3 = k\cdot k^2 \geq 4k^2 > 3k^2$ and $k^3 = k^2 \cdot k \geq 4^2 \cdot k > 4k = 3k + k \geq 3k + 4 > 3k + 1.$ – Dave L. Renfro Jul 11 '23 at 14:00
  • @DaveL.Renfro Yes this works +1, We can also use the method of Polynomial long division or Synthetic division, - I added to the answer . – lone student Jul 12 '23 at 06:45
  • @lone student: It's interesting how many ways this can be done. I think that's due to how strong the inequality is. When I was first working out what I gave in my previous comment, I expected to see "close calls" when establishing the inequalities (if I could even establish an appropriate inequality for each left-side $k^3$ separately, which I didn't know initially), but I found myself making extremely generous replacements (e.g. replacing $4^2 = 16$ with $3$ and replacing $4$ with $1).$ Before simply expanding $(k+1)^3,$ (continued) – Dave L. Renfro Jul 12 '23 at 09:37
  • I very briefly considered $\left(\frac{k+1}{k}\right)^3 = \left(1 + \frac{1}{k}\right)^3,$ which I suspected was less than $3,$ since $\left(1 + \frac{1}{k}\right)^k \rightarrow e = 2.71828\ldots$ as $k \rightarrow \infty$ and the limit is approached monotonically from below (see here), but I felt that trying make use of calculations involving this sequence was going to be much more difficult than I wanted to deal with, so I immediately dismissed that approach and tried expanding $(k+1)^3.$ – Dave L. Renfro Jul 12 '23 at 09:37
  • @DaveL.Renfro Nice and clever observations, indeed ! By the way, I've expanded my answer by also adding a "less" computational way : $$\begin{align}&\thinspace\thinspace\thinspace3^{k+1}-(k+1)^3\ =&\thinspace\thinspace\thinspace3(a+k^3)-(k+1)^3\ >&\thinspace\thinspace\thinspace 3k^3-(k+1)^3\ >&\thinspace\thinspace\thinspace 3k^3-\left(k+\frac k5\right)^3\ =&\thinspace\thinspace\thinspace \frac{159}{125}k^3\thinspace >\thinspace 0\end{align}$$ – lone student Jul 12 '23 at 16:53
  • 1
    @lone student: It seems that my wording "which I suspected" should have been "which I knew" (but I gave this approach almost no thought yesterday, and I suppose I was too sleep-deprived in my 8 hours-ago-comment to notice), since knowing for each integer $k>3$ that $\left(1 + \frac{1}{k}\right)^k < e < 3$ (although showing this from scratch is harder than what we want to show) and $\left(1 + \frac{1}{k}\right)^3 < \left(1 + \frac{1}{k}\right)^k,$ it follows that for each integer $k>3$ we have $\left(1 + \frac{1}{k}\right)^3 < 3.$ – Dave L. Renfro Jul 12 '23 at 17:53