# Local & global extrema

#### mathmari

##### Well-known member
MHB Site Helper
Hey !! Let $f:\mathbb{R}^2\rightarrow \mathbb{R}^2$, $f(x,y)=\sin^2(x)\cdot \cos^2(y)$.

- Show that $f$ has at $\left (\frac{\pi}{2}, 0\right )$ a strictly local maximum and that is also a global maximum.

- Determine all points at which $f$ gets its global minimum.

I have sone the following :

It holds that $|\sin (x)|\leq 1$ so $\sin^2(x)\leq 1$ and also that $|\cos (x)|\leq 1$ so $\cos^2(x)\leq 1$.

From that we get that $f(x,y)\leq 1$, that means that $1$ is a maximum of $f$.

At the point $\left (\frac{\pi}{2}, 0\right )$ the function gets the value $1$, that means that $f$ has at $\left (\frac{\pi}{2}, 0\right )$ a strictly local maximum.

Is that correct so far? #### Klaas van Aarsen

##### MHB Seeker
Staff member
It holds that $|\sin (x)|\leq 1$ so $\sin^2(x)\leq 1$ and also that $|\cos (x)|\leq 1$ so $\cos^2(x)\leq 1$.

From that we get that $f(x,y)\leq 1$, that means that $1$ is a maximum of $f$.
Hey mathmari !!

At this stage we don't know yet if $1$ is an actual maximum of the function.
We only know that $1$ is an upper bound of the function. At the point $\left (\frac{\pi}{2}, 0\right )$ the function gets the value $1$, that means that $f$ has at $\left (\frac{\pi}{2}, 0\right )$ a strictly local maximum.
We now have a point where we achieve the upper bound, which implies that $1$ is indeed a maximum of the function.
However, at this stage we don't know yet, if it is a strictly local maximum. Last edited:
• mathmari

#### mathmari

##### Well-known member
MHB Site Helper
However, at this stage we don't know yet, if it is a strictly local maximum. Do we have to show that in the neighbourhood of the point the function achieves only at this point the value $1$ ? Or how do we show that? #### Klaas van Aarsen

##### MHB Seeker
Staff member
Do we have to show that in the neighbourhood of the point the function achieves only at this point the value $1$ ? Or how do we show that?
Yes. That is what it means that a local maximum is strict. The function $\sin x$ has a strict local maximum at $x=\frac\pi 2$, doesn't it? • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
The function $\sin x$ has a strict local maximum at $x=\frac\pi 2$, doesn't it? So we have that for each other $x$ it holds that $\sin(x)<1$ and so we get that $\sin^2 (x)\cdot \cos^2 (y)<1$ for $x\neq \frac{\pi}{2}$, right?

So can we say that this means that at $\left (\frac{\pi}{2}, 0\right )$ the function has a strictly local maximum? #### Klaas van Aarsen

##### MHB Seeker
Staff member
So we have that for each other $x$ it holds that $\sin(x)<1$ and so we get that $\sin^2 (x)\cdot \cos^2 (y)<1$ for $x\neq \frac{\pi}{2}$, right?
Only in a small enough neighborhood of $(\frac\pi 2,0)$. So can we say that this means that at $\left (\frac{\pi}{2}, 0\right )$ the function has a strictly local maximum?
We still have the case that we are in a small enough neighborhood with $x= \frac{\pi}{2}$.
Can we deduce in that case as well that the function value is strictly less than $1$? • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
Only in a small enough neighborhood of $(\frac\pi 2,0)$. We still have the case that we are in a small enough neighborhood with $x= \frac{\pi}{2}$.
Can we deduce in that case as well that the function value is strictly less than $1$? In a small neighborhood of $0$ the function $\cos^2(y)$ is only $1$ at $y=0$ and in a small neighborhood of $\frac{\pi}{2}$ the function $\sin^2(x)$ is only $1$ at $x=\frac{\pi}{2}$, that means that in a small neighborhood of $(\frac\pi 2,0)$ the function $f$ is less than $1$.

Ia that correct so far? #### Klaas van Aarsen

##### MHB Seeker
Staff member
In a small neighborhood of $0$ the function $\cos^2(y)$ is only $1$ at $y=0$ and in a small neighborhood of $\frac{\pi}{2}$ the function $\sin^2(x)$ is only $1$ at $x=\frac{\pi}{2}$, that means that in a small neighborhood of $(\frac\pi 2,0)$ the function $f$ is less than $1$.

Is that correct so far?
Yep. • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
Yep. So we get that $f$ has at this point a strictly local maximum , right?

To show that this is also a global maximum do we say that it cannot be greater than $1$ ?

Also can we just mention the parts with the small neighborhoods at #7 or do we have to prove that? #### Klaas van Aarsen

##### MHB Seeker
Staff member
So we get that $f$ has at this point a strictly local maximum , right?

To show that this is also a global maximum do we say that it cannot be greater than $1$ ?
Indeed.
Suppose it was not a global maximum. Then there would have to be another point where $f$ is greater than $1$.
But we already established that $1$ is an upper bound, leading to a contradiction.
So it is indeed also a global maximum. Also can we just mention the parts with the small neighborhoods at #7 or do we have to prove that?
The functions $\sin$ and $\cos$ are well known to have a strict local maximum at $\frac\pi 2$ respectively $0$.
So they are less than $1$ in a sufficiently small neighborhood around those points.
Consequently their squares and product are also less than $1$ in a sufficiently small neighborhood.
I believe that is sufficient to serve as proof. • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
Indeed.
Suppose it was not a global maximum. Then there would have to be another point where $f$ is greater than $1$.
But we already established that $1$ is an upper bound, leading to a contradiction.
So it is indeed also a global maximum. The functions $\sin$ and $\cos$ are well known to have a strict local maximum at $\frac\pi 2$ respectively $0$.
So they are less than $1$ in a sufficiently small neighborhood around those points.
Consequently their squares and product are also less than $1$ in a sufficiently small neighborhood.
I believe that is sufficient to serve as proof. Ok! I see!! At the second part,where we have to determine all points at which $f$ gets its global minimum, do we have to do the same as in the first part just for minimum instead of maximum ? #### Klaas van Aarsen

##### MHB Seeker
Staff member
At the second part,where we have to determine all points at which $f$ gets its global minimum, do we have to do the same as in the first part just for minimum instead of maximum ?
Yep. • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
Yep. Ah the minimum in this case is $0$ because of the square, right? #### Klaas van Aarsen

##### MHB Seeker
Staff member
Ah the minimum in this case is $0$ because of the square, right?
Indeed. • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
Indeed. So we have that $\sin^2(x)\cdot \cos^2(y)\geq0$ and at all points with $x=n\pi$ and all points with $y=(2k-1)\frac{\pi}{2}$ the function achieves that minimum, which is a global minimum because there is no other lower value of the function.

Is that correct and complete? Or do we have to prove something more? #### Klaas van Aarsen

##### MHB Seeker
Staff member
So we have that $\sin^2(x)\cdot \cos^2(y)\geq0$ and at all points with $x=n\pi$ and all points with $y=(2k-1)\frac{\pi}{2}$ the function achieves that minimum, which is a global minimum because there is no other lower value of the function.

Is that correct and complete? Or do we have to prove something more?
Correct and complete. • mathmari

#### mathmari

##### Well-known member
MHB Site Helper
Correct and complete. Great!! Thank you very much!! • Klaas van Aarsen