1

How to prove that $\frac{a+b}{c} \leq \sqrt2$, where a, b and c are the sides of a right angled triangle and c is the hypotenuse?

My attempt:

$a+b>c$ gives $\frac{a+b}{c}$ > 1

$a<b$ and $b<c$ , adding both inequalities we get $a+b<2c$ or $\frac{a+b}{c}$<2

so for a general triangle we have the upper and lower limits of the required expression.

Now, I am confused on how to proceed further for the specific case of right angled triangle, as far as I can understand is that we need to manipulate the expression $a^2+b^2=c^2$, but can't make much progress in here

Red Five
  • 3,164
Vasu Gupta
  • 1,090
  • 3
  • 12
  • 1
    Your question should be clear without the title. After the title has drawn someone's attention to the question by giving a good description, its purpose is done. The title is not the first sentence of your question, so make sure that the question body does not rely on specific information in the title. – Martin R May 23 '24 at 04:37
  • Are you familiar with AM-GM? How about the Power mean inequalities? – Calvin Lin May 23 '24 at 04:39
  • Hi @CalvinLin, I am familiar with AM-GM inequality , not for other one you mentioned – Vasu Gupta May 23 '24 at 04:40
  • So you want to show that $(a+b)^2 \leq 2c^2$ right? What can we do from here? – Calvin Lin May 23 '24 at 04:42
  • Cauchy-Schwarz inequality ... – Martin R May 23 '24 at 04:43
  • 1
    We can expand $(a+b)^2$ and then substitute $a^2+b^2$ as $c^2$ ? – Vasu Gupta May 23 '24 at 04:43
  • We shall get $c^2 + 2ab <= 2c^2$ or $c^2>=2ab$ or $ab<=\frac{c^2}{2}$... after this if we apply AM>=GM inequality, we get $\frac{a+b}{2}$>= sqrt(ab) or $\frac{a+b}{2}$>= $\frac{c}{sqrt(2)}$ ... I have proved the opposite of what was required @CalvinLin :( – Vasu Gupta May 23 '24 at 04:50
  • How did you do that very last part? You seem to be using $ \sqrt{ab} \geq \frac{c}{\sqrt{2}}$, when in actuality it's the other way around (so we can't apply it here). $\quad$ You should be clear with what you're wanting to show, vs what you know. EG You want to show that $c^2 \geq 2ab$ vs you know that $ (a+b) /2 \geq \sqrt{ab}$. $\quad$ I'd encourage you to try that again. Reducing the number of variables could help, so I'd suggest replacing $ c^2 = a^2 b^2$ which uses the right-triangle condition. – Calvin Lin May 23 '24 at 04:57
  • See https://math.stackexchange.com/q/240034/42969 and https://math.stackexchange.com/q/3123485/42969 – Martin R May 23 '24 at 05:16

4 Answers4

2

By AM-GM $$a^2+b^2≥2ab$$ Add $a^2+b^2$ to both sides. It's a positive value. $$2(a^2+b^2)≥a^2+b^2+2ab \implies 2c^2≥(a+b)²$$ Square root both sides and get $$\sqrt 2 c≥a+b$$

Gwen
  • 3,910
  • 6
  • 19
1

Consider a square with side length $a+b$ and an inner square with side length $c$ such that the inner square is incident to the outer square.

Inner square incident to outer square

We are now going to prove the following lemma:

The smallest possible inner square area to outer square area ratio happens when the blue triangles are isosceles ($a=b$).

Proof 1 (simpler, credit: Calvin Lin):

Notice that the area of a square is equal to $\frac{1}{2}d^2$, where $d$ is the diagonal. It is relatively easy to show that the smallest that the diagonal of the inner square can be is when $d = a+b$. In all other cases, $d > a+b$ by Pythagorean.

Now it is easy to show that the smallest possible inner square will be half the area of the outer square.

$$ \frac{1}{2} (a+b)^2 \le c^2 \\\\ \frac{(a+b)^2}{c^2} \le 2 \\\\ \frac{a+b}{c} \le \sqrt{2} $$

Proof 2 (trigonometry):

As the area of a right triangle with hypotenuse $1$ is

$$\frac{\sin\theta\cos\theta}{2}=\frac{\sin 2\theta}{4}$$

then this is a maximum when $\theta=45^\circ,\;\; 0\le\theta\le45^\circ$. This implies the maximum occurs when $a=b$.

Knowing this makes it easy to calculate the area of each blue triangle (at the maximum) as $\frac{(a+b)^2}{8}$. There are $4$ such triangles, so they make up a total area of $\frac{(a+b)^2}{2}$, which is half of the outer square area (and therefore equal to the inner square area).

$$\frac{(a+b)^2}{c^2}\le2$$

Motivation:

The biggest spark of intuition comes from the $\sqrt{2}$ value, as it could have something to do with the side length ratios of a 45-45-90 degree triangle.

Playing with the values and testing some triangles might lead you to the hypothesis that the maximum probably occurs when $a=b$, further confirming that the 45-45-90 triangle is related.

Visually, this makes sense too: the "thinner" the right triangle, the closer the legs are to the hypotenuse. The isosceles right triangle pulls the legs furthest away from the hypotenuse.

Once the goal is more clear, now it is just a matter of how to prove such an idea. Here, it helps to see that when the maximum of $\frac{a+b}{c}$ is attained, the area of the right triangle is also maximized (relative to fixing $a+b$ or $c$). Combining that with some basic algebra rearrangement could therefore lead to this proof to make these intuitions rigorous.

zhuli
  • 2,768
JMP
  • 22,637
  • 2
    How does the scaling and rotation lead to the obvious conclusion that the relative area is $\leq 2$? $\quad$ FWIW I like your idea and there's a way to prove it, but I don't think your reasoning is as clear/obvious as you stated (esp for someone who's struggling with the inequalities) – Calvin Lin May 23 '24 at 04:59
  • @CalvinLin; fixed. – JMP May 23 '24 at 06:04
  • 2
    My approach was to observe that the area of a square is equal to half of the square of the diagonal. Since the diagonal of the inner square is more then the side length, hence $c^2 = $ inner square area $ = \frac{1}{2} d^2 \geq \frac{1}{2} (a+b)^2 $. This gives us the desired inequality. – Calvin Lin May 23 '24 at 08:42
  • I think this is a very natural approach to the problem due to the intuition that the $\sqrt{2}$ induces, and Calvin's approach to the relative area part of the problem is much simpler and cleaner.

    I'm going to make quite a heavy edit to the answer here because I don't think the changes I'm making warrant a new answer, but I also believe the solution deserves it. If anyone of interest disagrees, please feel free to rollback my edits.

    – zhuli May 23 '24 at 12:32
1

Using $a^2+b^2=c^2$ we have $$ 0\leq(a-b)^2 \implies 2ab\leq c^2 \\ \implies a^2+b^2+2ab\leq 2c^2 \implies a+b\leq \sqrt{2}c. $$

Simple version of the solution by @Oppenheimer.

Steen82
  • 1,258
0

As you have pointed out,

a + b $>$ c

to prove your question,

a + b $\leq$ $\sqrt{2}c$

square both the sides,

$a^{2}$ + $b^{2}$ + 2ab $\leq$ $2c^{2}$

since $a^{2} + b^{2} = c^{2}$,

it becomes,

$2ab$ $\leq$ $c^{2}$

Case - I, suppose $2ab = c^{2}$

substituting, $c^{2} = a^{2} + b^{2}$,

$( a - b )^{2} = 0$

this case is true when a = b,

Case - II, $2ab$ $\leq$ $c^{2}$ We shall prove this result by way of contradiction,

Assume the opposite,

$2ab$ $\geq$ $c^{2}$

suppose 2ab - k = $a^{2} + b^{2}$

where k is a positive integer,

-k = $( a - b )^{2}$

Which cannot be true, as the square of anything is positive ( Ignoring complex numbers )

Which completes our proof.