Proving A=B with product law.

Integrate

Junior Member
Joined
May 17, 2018
Messages
118
a=b.png


I know I am supposed to show my work before submitting a question but I for the life of me just don't even know how to absorb this question.


This was in rogawski's summary and I don't remember a single question like this in the limits chapter.


Can I get a nudge in the right direction?
 
I sometimes do proofs by parts (less and less as I mature mathematically).
Can you prove the statement if B=/ 0? You CAN you the Quotient Law.
Now assume B=0. Can you prove the theorem for this case.

For the record, blamocur stated an excellent way to prove the theorem. I am just offering another way of seeing things. Even if you can't prove one of my two cases, at least you get credit for the case you could prove.
 
Let's see if I can prove this by definition.
We are given that \(\displaystyle \lim_{x \to a}f(x) = A, \lim_{x \to a}g(x)=B\ and\ \lim_{x \to a}\dfrac{f(x)}{g(x)} = 1\)

\(\displaystyle So,\ given\ \epsilon >0,\ \exists\ \delta_1>0\ such\ that\ \mid f(x) - A \mid < \epsilon\ whenever\ 0< \mid x - a \mid \leq \delta_1 \)

\(\displaystyle and\ for\ the\ same\ \epsilon >0,\ \exists\ \delta_2>0\ such\ that\ \mid g(x) - B \mid < \epsilon\ whenever\ 0< \mid x - a \mid \leq \delta_2 \)

\(\displaystyle \lim_{x \to a}\dfrac{f(x)}{g(x)} = 1 \iff given\ any\ \epsilon>0, \exists\ \delta >0 such\ that\ \mid \dfrac {f(x)}{g(x)} - 1 \mid = \mid \dfrac {f(x)-g(x)}{g(x)} \mid< \epsilon\ whenever\ 0< \mid x - a \mid \leq \delta\)

Now if we take \(\displaystyle \delta\) smaller enough (at most \(\displaystyle \delta = min\){\(\displaystyle \delta_1, \delta_2\)}), g(x) will be with 1 of A. That is |g(x)-A|<1 (not the same 1 as the limit of f(x)/g(x))
Hence A-1<g(x)<A+1, so |g(x)| < max{|A-1|, |A+1|} = C

Now, \(\displaystyle \mid \dfrac {f(x)-g(x)}{C} \mid\ \leq\ \mid \dfrac {f(x)-g(x)}{g(x)} \mid < \epsilon\)

Putting this all together we have that given ϵ>0, ∃ δ>0 such that \(\displaystyle \mid f(x)-g(x) \mid\ < C\epsilon\ whenever\ 0< \mid x - a \mid \leq \delta\)

By definition of limit, this means that \(\displaystyle \lim_{x \to a}f(x)-g(x)=0,\ or\ that\ \lim_{x \to a}f(x)\ =\ \lim_{x \to a}g(x)\). Hence A=B.

I feel like a student right now by asking you nice helpers to please grade this proof, point out any any errors and how it can be improved?
Thanks!
 
Last edited:
Beer induced suggestion follows.
View attachment 34607


I know I am supposed to show my work before submitting a question but I for the life of me just don't even know how to absorb this question.


This was in rogawski's summary and I don't remember a single question like this in the limits chapter.


Can I get a nudge in the right direction?
It's an odd numbered exercise.
The solution is at the back of your textbook.
I suggest you look it up, take a screenshot, and post it here if you like.
 
@Steven G

Analysis is not my strong suit. I studied it almost sixty years ago, did not like it, and never had any practical use for it. So my criticism may be inane.

One clause strikes me as outright wrong, but I suspect it is irrelevant. "Not the same 1 as the limit of f(x)/g(x)" makes no sense. One is a number, and many things may equal 1. But one does not come in flavors.

Another thing strikes me as unclear. I can see that C is necessarily positive. But is C necessarily [imath]\le[/imath] 1? What if g(x) is negative in the given interval. In short I see what you are doing, but it works only if C does not exceed 1. And the logic seems to assume that [imath]g(x) = |g(x)|[/imath]. As I said, my comments may be inane
 
@Steven G

Analysis is not my strong suit. I studied it almost sixty years ago, did not like it, and never had any practical use for it. So my criticism may be inane.

One clause strikes me as outright wrong, but I suspect it is irrelevant. "Not the same 1 as the limit of f(x)/g(x)" makes no sense. One is a number, and many things may equal 1. But one does not come in flavors.

Another thing strikes me as unclear. I can see that C is necessarily positive. But is C necessarily [imath]\le[/imath] 1? What if g(x) is negative in the given interval. In short I see what you are doing, but it works only if C does not exceed 1. And the logic seems to assume that [imath]g(x) = |g(x)|[/imath]. As I said, my comments may be inane
Thanks for your comment.

As for the issue with 1--I am just trying to say that I did not pick 1 because one of the limits equaled 1. It was just a coincidence.

Since Cϵ is just another positive number, why do you think that Cϵ>ϵ can't work? I think that is wrong for the following reason. Why can't my epsilon be ϵ/C, so C times epsilon would be, well ϵ. The more I think about it, it seems fine. I too am not too strong with analysis so I might be wrong and would like to read your reply to this.

I do not see where I am saying that |g(x)| = g(x). Can you state where/why you think that I am using that fact?

Thanks!
Steven
 
Thanks for your comment.

As for the issue with 1--I am just trying to say that I did not pick 1 because one of the limits equaled 1. It was just a coincidence.
OK. I get it.
Since Cϵ is just another positive number, why do you think that Cϵ>ϵ can't work? I think that is wrong for the following reason. Why can't my epsilon be ϵ/C, so C times epsilon would be, well ϵ.
I do not think it is in the spirit of this kind of proof to put any constraint on [imath]\epsilon[/imath], which is supposed to be an arbitrary positive number (if I remember any of this correctly). By definition, C is positive. Obviously, if C does not exceed 1, [imath]C\epsilon \le \epsilon[/imath][math][/math] But there is no guarantee that either [imath]\delta_1[/imath] or [imath]\delta_2[/imath] are below 1. That will depend on [imath]\epsilon[/imath]. Perhaps all you need to do to address this is to define

[math]0 < |g(x)| < |B - 1| < C = \min \{ \delta_1, \delta_2, 1\}.[/math]
I think you meant B rather than A. In any case since you are defining C, you can force it not to exceed 1.

The more I think about it, it seems fine. I too am not too strong with analysis so I might be wrong and would like to read your reply to this.

I do not see where I am saying that |g(x)| = g(x). Can you state where/why you think that I am using that fact?
This is where I I got confused. You said that you were doing this without cases. But I did not see how you addressed the changes involved in an inequality when dividing it by a possibly negative number coupled with absolute values. In fact, it looks as though the problem arises when g(x) > 0. I have now worked out (hopefully without a blunder) what was worrying my intuition.

[math] 0 < |g(x)| < C \implies\\ -C < g(x) < 0 \text { if } g(x) < 0 \text { or } 0 < g(x) < C \text { if } g(x) > 0 \implies\\ -1 < \dfrac{g(x)}{C} < 0 \text { if } g(x) < 0 \text { or } 0 < \dfrac{g(x)}{C} < 1 \text { if } g(x) > 0 \implies\\ - \dfrac{1}{g(x)} > \dfrac{1}{C} > 0 \text { if } g(x) < 0 \text { or } 0 < \dfrac{1}{g(x)} < \dfrac{1}{C} \text { if } g(x) > 0 \implies\\ \left |- \dfrac{1}{g(x)} \right | > \left | \dfrac{1}{C} \right | > 0 \text { if } g(x) < 0 \text { or } 0 < \left | \dfrac{1}{g(x)} \right | < \left |\dfrac{1}{C} \right | \text { if } g(x) > 0 \implies\\ \left | \dfrac{1}{g(x)} \right | > \left | \dfrac{1}{C} \right | > 0 \text { if } g(x) < 0 \text { or } 0 < \left | \dfrac{1}{g(x)} \right | < \left |\dfrac{1}{C} \right | \text { if } g(x) > 0 \implies\\ \left | \dfrac{f(x) - g(x)}{g(x)} \right | \ge \left | \dfrac{f(x) - g(x)}{C} \right | > 0 \text { if } g(x) < 0 \text { or }\\ 0 < \left | \dfrac{f(x) - g(x)}{g(x)} \right | \le \left |\dfrac{f(x) - g(x)}{C} \right | \text { if } g(x) > 0. [/math]
So I did not see how you arrived at [imath]\left | \dfrac{f(x) - g(x)}{g(x)} \right | \ge \left | \dfrac{f(x) - g(x)}{C} \right | > 0[/imath] regardless of the sign of g(x). Maybe I made a mistake above. And maybe, if I did not make a mistake, there is a way to fix your proof. You asked for a review; I did not promise you a competent one. Just trying to help a friend.
 
Jeff,
Epsilon can be any positive number. Fair enough, we want epsilon to be small. Even if C>1 (or C>>1), Cϵ can still be as small as you like by adjusting ϵ. Since epsilon is any positive number, epsilon can be say 5ϵ', where ϵ' is any positive real number.

I need to look at the fact that I used A. I need to also take into account both A and B to compute C.

Something just came up and I can't look into for a few hours.

Thanks for you time and input.

Steve
 
Thanks for your comment.

As for the issue with 1--I am just trying to say that I did not pick 1 because one of the limits equaled 1. It was just a coincidence.

Since Cϵ is just another positive number, why do you think that Cϵ>ϵ can't work? I think that is wrong for the following reason. Why can't my epsilon be ϵ/C, so C times epsilon would be, well ϵ. The more I think about it, it seems fine. I too am not too strong with analysis so I might be wrong and would like to read your reply to this.

I do not see where I am saying that |g(x)| = g(x). Can you state where/why you think that I am using that fact?

Thanks!
Steven
See proof #4.
 
Jeff,
Epsilon can be any positive number. Fair enough, we want epsilon to be small. Even if C>1 (or C>>1), Cϵ can still be as small as you like by adjusting ϵ. Since epsilon is any positive number, epsilon can be say 5ϵ', where ϵ' is any positive real number.

I need to look at the fact that I used A. I need to also take into account both A and B to compute C.

Something just came up and I can't look into for a few hours.

Thanks for you time and input.

Steve
I still think that you cannot use something you need in the establishment of delta to set epsilon. My recollection is that epsilon is a constant that is totally arbitrary. You cannot make it suit your convenience. But as I said, my recollection of analysis is hazy.
 
So like. How did you know to do this?
Two common techniques are to multiply by one in a helpful form or to add zero in a helpful form.

[math]\text { For all } x \text { such that } f(x) \text { is defined and } g(x) \ne 0,\\ f(x) = f(x) * 1 = f(x) * \dfrac{g(x)}{g(x)} = g(x) * \dfrac{f(x)}{g(x)}.[/math]
The creative part is finding which form is helpful
 
Top