# TrigonometryUsing of Jensen's Inequality to prove cosA+cosB+cosC less than or equal to 3/2 (where A+B+C=pi)

#### anemone

##### MHB POTW Director
Staff member
Hi,
Given $A+B+C=\pi$, I need to prove $cosA+cosB+cosC\leq \frac{3}{2}$.

I wish to ask if my following reasoning is correct.
First, I think of the case where A and B are acute angles, then I can use the Jensen's Inequality to show that the following is true.
$cos\frac{A+B}{2}\geq \frac{cosA+cosB}{2}$
Carrying on with the working, I get
$sin\frac{C}{2}\geq \frac{cosA+cosB}{2}$
$2sin\frac{C}{2}\geq cosA+cosB$
$cosA+cosB\leq 2sin\frac{C}{2}$
$cosA+cosB+cosC\leq 2sin\frac{C}{2}+cosC$
$cosA+cosB+cosC\leq 2sin\frac{C}{2}+1-2sin^2C$
Completing square the RHS to obtain
$cosA+cosB+cosC\leq -2(sin\frac{C}{2}-\frac{1}{2})^2+\frac{3}{2}$

Now, it's obvious to see that $cosA+cosB+cosC\leq \frac{3}{2}$

My question is, can I solve the question by thinking A and B are acute angles and ignore the angle C right from the start so that I can let f(x)=cosx and notice that the curve of f(x)=cos x in the interval $x\in (0,\frac{\pi}{2})$ take the convex shape which in turn I can apply the Jensen's inequality without a problem?

Thanks.

#### Opalg

##### MHB Oldtimer
Staff member
My question is, can I solve the question by thinking A and B are acute angles and ignore the angle C right from the start so that I can let f(x)=cosx and notice that the curve of f(x)=cos x in the interval $x\in (0,\frac{\pi}{2})$ take the convex shape which in turn I can apply the Jensen's inequality without a problem?
Yes: a triangle can have at most one obtuse angle, and the other two (acute) angles must necessarily be adjacent. Without loss of generality (favourite math phrase), take them to be A and B.

And BTW that is a very neat proof!

• anemone

#### anemone

##### MHB POTW Director
Staff member
Yes: a triangle can have at most one obtuse angle, and the other two (acute) angles must necessarily be adjacent. Without loss of generality (favourite math phrase), take them to be A and B.
Thanks, Opalg.
And I think 'obviously' is also one of the favourite math phrase too!
But unfortunately that is a phrase to which I considered not so true and annoying some (if not most) of the time. And BTW that is a very neat proof! 