3

These are just the examples from my textbook, but I don't think it did not explain well.

One of the problem was to prove the inequality

$$n<2^n$$

for all integers $n$.

I understand we assume if $P(n)$ is true, $P(k+1)$ is true.

Thus $k+1< 2^{k+1}$

But in the text book, they added one more line that I don't understand:

$$k+1<2^k+1 \le 2^k+2^k=2\cdot2^k=2^{k+1}$$

I understand $$k+1<2^{k+1}$$

But how do I know $2^k+1$ is larger than $k+1$, and smaller than $2^{k+1}$?

I have one another example just in case,

proving $2^n<n!$ using induction.

$$2^{k+1}=2\cdot 2^k<2\cdot k!<(k+1)k!=(k+1)!$$

I understand we should start with $2^{k+1}<(k+1)!$, but

Where did $2\cdot k!$ come from? Textbook says it came from inductive hypothesis, but I still don't get it.

Please advise.

Zain Patel
  • 17,200
  • 5
  • 29
  • 59
Minjae
  • 207
  • Do you remember that the first step in an inductive proof is testing the base case (that is, verifying the hypothesis holds for some starting value, usually $1$, the first natural number, sometimes $0$)? – CiaPan Jul 04 '15 at 21:40
  • Yes. That comes out when $k=1$, $1<2^1$, right? – Minjae Jul 04 '15 at 21:48
  • Right. Also for $k=0$ there is $k = 0 <1 =2^0 = 2^k$. That might be useful if you decided to run the induction into negative numbers, too (the exercise says 'for all integers', not 'for all positive integers'). – CiaPan Jul 04 '15 at 21:55

3 Answers3

2

We assume that $$k < 2^k$$ add one to both sides, we get $$k+1 < 2^k + 1$$

But we also know that $2^k + 1 < 2^k + 2^k$ since $1 < 2^k$ for every natural $k$. This means we have $$2^k + 1 < 2^{k+1}$$ which completes our induction.

So, to answer your question, you need to know that $2^k > 1$ for every natural k, so that $2^k + 1 < 2^{k} + 2^{k}$. This technique comes up often in inductions using inequalities.

For your next example, assume $$2^k < k!$$ We have that $$(k+1)! = (k+1)k!$$

But we assumed that $2^k < k!$ so we have $$2\cdot 2^k < (k+1)k!$$ (I'm making the assumption that $n \geq 4$) and hence $$2^{k+1} < (k+1)!$$ completing our induction.

Zain Patel
  • 17,200
  • 5
  • 29
  • 59
2

You need to prove that $P(n) \Rightarrow P(n+1)$. So you take an arbitrary $n$ and suppose P(n) true. This gives you that

$$n < 2^n$$

You add one to each side, and you get

$$n+1<2^n+1$$

Now, you know that $\forall n \in \mathbb{N}, 2^n \geq 1$, hence

$$n+1<2^n+1 \leq 2^n+2^n = 2^{n+1}$$

And this is exactly P(n+1)

Tryss
  • 14,490
  • 20
  • 35
2

But how do I know $2^k +1$ is larger than $k+1$

Because your inductive step starts from the assumption that the given statement holds for $k$, which means $$k < 2^k$$ When added $1$ to both sides it results in $$k+1< 2^k +1$$

CiaPan
  • 13,884