monotone functions

**alexandros** · May 3rd 2011, 02:56 AM

Given that:

1) f,g,h are real nalued funtions ,with f strictly monotone and the other two monotone

2) g(f(x)) = h(f(x)) =x for all ,x ,belonging to the real Nos

3) the domain of all 3 being the real Nos

Prove that:

g=h for all ,x, belonging to the real Nos

**HallsofIvy** · May 3rd 2011, 07:19 AM

I would be inclined to an "indirect" proof. Suppose g is not equal to h. Then there exist y such that [itex]g(y)\ne h(y)[/itex]. Does there exist x such that f(x)= y?

**alexandros** · May 3rd 2011, 02:39 PM

Originally Posted by HallsofIvy

I would be inclined to an "indirect" proof. Suppose g is not equal to h. Then there exist y such that [itex]g(y)\ne h(y)[/itex]. Does there exist x such that f(x)= y?

To prove that such ,x , exists we must prove that f is onto.

But how do we prove that f is onto?

**SlipEternal** · May 3rd 2011, 07:56 PM

Well, you have $g(f(x))=x=h(f(x))$ . So, what is: $g(f(g(f(x))))$ ? What about $g(f(g(f(g(f(x))))))$ ? In other words, what do you know about $g$ and $h$ ?

**alexandros** · May 4th 2011, 12:25 AM

Originally Posted by SlipEternal

Well, you have $g(f(x))=x=h(f(x))$ . So, what is: $g(f(g(f(x))))$ ? What about $g(f(g(f(g(f(x))))))$ ? In other words, what do you know about $g$ and $h$ ?

This has nothing to do with f being onto,besides g(f(g(f(x)))) =x

**SlipEternal** · May 4th 2011, 01:36 AM

Originally Posted by alexandros

This has nothing to do with f being onto,besides g(f(g(f(x)))) =x

Sorry, I was hoping you would see that $f$ brings $x$ to $f(x)$ and both $g,h$ bring it back... so when restricted to the image of $f$ , both $g,h$ act as inverse functions. So, try to determine properties for the image of $f$ . For instance, assume $sup\{f(x): x \in \mathbb{R}\}=0$ What would $\lim_{t \to 0}{g(t)}$ be? What about $g(1)$ ? You don't need $f$ to be onto. In fact, it can be extremely discontinuous, and by looking at the properties of its image, then the properties of its inverse, you should be able to quickly arrive at your conclusion.

**alexandros** · May 4th 2011, 04:10 PM

Originally Posted by SlipEternal

Sorry, I was hoping you would see that $f$ brings $x$ to $f(x)$ and both $g,h$ bring it back... so when restricted to the image of $f$ , both $g,h$ act as inverse functions. So, try to determine properties for the image of $f$ . For instance, assume $sup\{f(x): x \in \mathbb{R}\}=0$ What would $\lim_{t \to 0}{g(t)}$ be? What about $g(1)$ ? You don't need $f$ to be onto. In fact, it can be extremely discontinuous, and by looking at the properties of its image, then the properties of its inverse, you should be able to quickly arrive at your conclusion.

Listen,that problem bothered me for a couple of days.And i am nearly fed up with trying any more .

So if you have a solution at hand ,please ,if you wish that is, let me have it ,or at least let me have the skeleton of that solution and help me fill in the details

**SlipEternal** · May 4th 2011, 06:15 PM

Without loss of generality, assume that $f(x)$ is strictly monotone increasing. Assume that it is also bounded above and $s=\sup{\{f(x): x \in \mathbb{R}\}}$ . Therefore, as $x \to \infty, f(x) \to s, g(f(x)) \to \infty$ . Therefore, $\lim_{t \to s}g(t)=\infty$ . Because g is monotone, it must be monotone increasing, so $g(s+1) \ge g(s)$ . However, $\infty \notin \mathbb{R}$ , so it cannot be that $f(x)$ is bounded above. By the same argument for the infimum, it cannot be bounded below. Therefore, it is unbounded. If it is continuous, then it is onto.

Assume that $f$ is discontinuous at some $x$ . Then, consider $a=\lim_{t \to x^-}f(t)$ and $b=\lim_{t \to x^+}f(t)$ . It must be that $g(a)=x=g(b)$ , so for all values $t \in [a,b], g(t)=x$ .

Apply the same to h, and you are done. You will likely need to be more careful with notation than I have been, but this is the skeleton of the proof.

**alexandros** · May 9th 2011, 02:48 PM

Originally Posted by SlipEternal

Assume that $f$ is discontinuous at some $x$ . Then, consider $a=\lim_{t \to x^-}f(t)$ and $b=\lim_{t \to x^+}f(t)$ .

.

Does this statement result by negating the following statement:?

f is continiuous at R iff [for all xεR $\lim_{t\to x^-}f(t) = f(x)= \lim_{t\to x^+} f(t)]$ ??

**SlipEternal** · May 10th 2011, 10:34 AM

Yes, that is an apt explanation.

**alexandros** · May 12th 2011, 06:21 AM

Now,why g(a)=g(b)??

**SlipEternal** · May 12th 2011, 11:07 AM

Because $f$ cannot be everywhere discontinuous (as it would be unbounded on every interval). Therefore, its set of discontinuities must be denumerable (for more on this, you might want to investigate the image of $[a,b]$ under $f$ where $[a,b]$ is any interval. Try to make every point--or even an uncountable number of points--in the interval discontinuous, and you will find that because it is strictly monotone, its image must be unbounded.)

So, this implies that $f$ is continuous between each and every one of its discontinuity points. $\lim{t\to x^-}f(t)=a$ implies that for some $r>0, f$ is continuous on the half-open interval $[x-r,x)$ . So, continuity on that interval allows you to find points arbitrarily close to $x$ such that their image is arbitrarily close to $a$ . Applying the same thought process for $b$ , implies that $g(a)=g(f(x))=x=g(b)$ .

**alexandros** · May 12th 2011, 04:33 PM

O.k suppose g(a) =g(b) is that a contradiction??

**SlipEternal** · May 12th 2011, 04:41 PM

No contradiction, because $g$ is monotone, not strictly monotone. So if it is monotone increasing (as assumed), then $a<b \Rightarrow g(a) \le g(b)$

Thread: monotone functions

LinkBack

Thread Tools

Display

monotone functions

Similar Math Help Forum Discussions

Monotone sequence - how can this be a monotone sequence if it bounces between numbers

Monotone Functions

sequences of eventually monotone functions

monotone

Is every strictly locally monotone sequence strictly monotone?

Search Tags