1. ## Induction Proof

I'm stuck on this proof. I'm not sure if I've done something wrong, or if I'm just not seeing the next step. Any help would be appreciated.

Let a1, a2, ..., an be positive real numbers such that a1*a2*...*an=1. Prove by induction that (1+a1)(1+a2)...(1+an) >= 2^n. The hint for the problem says to try a reduction by introducing another variable that replaces two specially chosen numbers from the sequence.

The base case is simple enough. On the induction step, I assume a1*a2*...*an=1 implies (1+a1)(1+a2)...(1+an) >= 2^n. Then I let z=an*an+1. Thus, a1*a2*...*an-1*z implies (1+a1)(1+a2)...(1+an-1)(1+z) >= 2^n. I then multiplied both sides of the inequality by 2, to get:
(1+a1)(1+a2)...(1+an-1)(1+z)(2) >= 2^(n+1)
At this point, I figured I could try to show that (1+an)(1+an+1)>=(1+z)*2, which would lead to (1+a1)(1+a2)...(1+an)(1+an+1 >= 2^(n+1), but that didn't quite seem to work out.

Any thoughts?

Thanks!

2. I can see no reason to use induction on this.
We know that $x^2 + y^2 \ge 2xy$ for each x&y.
Thus we see that $a > 0,\;b > 0\quad \Rightarrow \quad a + b \ge 2\sqrt {ab}$.
From which we have $\left( {\forall k} \right)\left[ {\left( {1 + a_k } \right) \ge 2\sqrt {a _k } } \right]$.
Put it all together: $\prod\limits_{k = 1}^n {\left( {1 + a_k } \right)} \ge \prod\limits_{k = 1}^n {2\sqrt {a _k } = 2^n \sqrt {\prod\limits_{k = 1}^n {\left( {a_k } \right)} } = 2^n }$.

3. Originally Posted by Plato
I can see no reason to use induction on this.
Unfortunately, the problem specifically asks for a proof by induction. Thank you for your help though. Any other thoughts?

4. Originally Posted by Altaica
Any other thoughts?
Yes indeed I have other thoughts.
For years I have served as an editor for mathematics contest problems as well as commercial test. In that trade this is known as a “too cute problem”. Whoever wrote it has a “cute induction” proof in his/hers brain. But why do it that way when there is a much more standard and usual way of doing it.

5. I hear you, but the point of the exercise is for us to become more familiar with induction, and that's what I'm trying to figure out here. I'm very rusty, and I've given it my best shot and I'm stuck. I've tried everything I can think of. I've stared at the problem for hours, and I really don't know where to go next. Even a nudge in the right direction would be very much appreciated.

6. Originally Posted by Plato
Yes indeed I have other thoughts.
For years I have served as an editor for mathematics contest problems as well as commercial test. In that trade this is known as a “too cute problem”. Whoever wrote it has a “cute induction” proof in his/hers brain. But why do it that way when there is a much more standard and usual way of doing it.
1. For practice of the technique

2. Why do we need N>=72 proofs of Pythagoras's theorem?

RonL

7. Thanks guys for your feedback. Does anyone have any thoughts about solving this by induction?

8. Originally Posted by Altaica
Does anyone have any thoughts about solving this by induction?
Frankly, I do not see any way that induction applies to this problem.
From the statement, for each n the set of positive numbers ${a_k}$ may have nothing to do with one another.
For example, for n=2 we may have the set $\left\{ {2,\frac{1}{2}} \right\}$ while for n=3 we could have the set $\left\{ {6,\frac{1}{2},\frac{1}{3}} \right\}$.
It is hard to see how one would employ the inductive step if the sets $\left\{ {a_k } \right\}_{k = 1}^n \,\& \,\left\{ {b_k } \right\}_{k = 1}^{n + 1}$ are not related in any necessary way.

Are you sure that you have given us the exact statement of the problem?

9. Here is the exact wording:
"Let a1, a2, ..., an be positive real numbers such that a1*a2*...*an=1, n is a natural number greater than 0. Show by PMI (proof by mathematical induction) that (1+a1)(1+a2)...(1+an) >= 2^n."

The hint for the problem says to try a reduction by introducing another variable that replaces two specially chosen numbers from the sequence. I figured this to mean that when we try the inductive step, we should reduce the problem to the nth case (which we assumed to be true), and then work from there. But, I got stuck shortly after that.