Ask your own question, for FREE!
Mathematics 8 Online
OpenStudy (anonymous):

Arithmetic Derivative

OpenStudy (anonymous):

@Kainui

OpenStudy (anonymous):

\[ k=ab \implies k' = a'b+ab' \]When \(k\) is prime, \(k'=1\).

OpenStudy (anonymous):

We established that \[ (k^n)' = k'nk^{n-1} \]

OpenStudy (anonymous):

For every prime \(p\) such that \(p|k\): \[ k' = k\sum_p\frac 1p \]

OpenStudy (anonymous):

Rules about primes shouldn't matter too much, generalizing to composite is better, in my opnion.

OpenStudy (kainui):

We also established a couple of fun facts: \[(p^p)' = p^p\] and \[(n!)' = n! \sum_{k=1}^n \frac{k'}{k}\]

OpenStudy (anonymous):

What would \( (-1)' \) be? It would extend us to integers.

OpenStudy (kainui):

I think we should think more generally than (-1)' and try to look at \[\large (e^{i \theta})'\]

OpenStudy (anonymous):

Well, \(-1\) is tough because a negative number by a positive number is negative.

OpenStudy (kainui):

Actually maybe we should consider looking at just simply fractions first, they seem like they might be fairly interesting. For instance, look at this pattern that begins to emerge when I write their derivatives out. \[\large 1' \ 2' \ 3' \ 4' \ 5' \ 6' \ 7' \ 8' \ 9' \ 10' \\ \large 0 \ \ 1 \ \ 4 \ \ \ 1 \ \ 5 \ \ 1 \ 12 \ \ 6 \ \ 7 \ \ 1\] It sort of begins to count 4, 5, 6, 7 with apparently a random 12 and some 1's in there while completely avoiding 2 and 3. Maybe this isn't really a pattern? But it might give us an idea of how to interpolate?

OpenStudy (anonymous):

Oh yeah, remember the formula: \[ \frac{a'}{a}+ \frac{b'}{b} = \frac{a'b+ab'}{ab} = \frac{c'}{c} \]

OpenStudy (anonymous):

\[ 1 \sim c\\ p \sim x\\ c \sim f(x) \]

OpenStudy (anonymous):

\[ \frac{(-2)'}{-2}+\frac{(-2)'}{-2} = \frac{-2(-2)' + -2(-2)'}{4} = \frac{-4(-2)'}{4} \]We should expect that \[ \frac{(-2)'}{-2}+\frac{(-2)'}{-2} = \frac{4'}{4} \]This iimplies \[ -4(-2)' = 4' \]

OpenStudy (anonymous):

This implies that \((-2)' = -1\)

OpenStudy (kainui):

\[1 = (-1)(-1)\] implies \[0 = (-1)'(-1)+(-1)(-1)'\]\[0=-2(-1)'\]\[0=(-1)'\]

OpenStudy (anonymous):

\[ (-a)' = (-1)'a +(-1)a' = -a' \]

OpenStudy (kainui):

Yeah you beat me to it haha.

OpenStudy (anonymous):

Next is:\[ \left(\frac{1}{a}\right)' \]I know the answer, but I wonder how we can get it using our rules.

OpenStudy (kainui):

Try something like 4 = 12/3

OpenStudy (kainui):

it should be consistent.

OpenStudy (anonymous):

\[ (a^{-1})' = -a'a^{-2} \]

OpenStudy (anonymous):

\[ \left(\frac 1a\right)' = -\frac{a'}{a^2} \]

OpenStudy (kainui):

\[1= a *a^{-1}\]\[0 = (a)'a^{-1}+a(a^{-1})'\]

OpenStudy (anonymous):

\[ \left(\frac ab\right)' = (ab^{-1})' = a'b^{-1} + a(b^{-1})' \]

OpenStudy (kainui):

Yeah, it's all consistent with what we had before with (k^n)' like you wrote at the beginning.

OpenStudy (anonymous):

\[ = \frac{a'}{b} + \frac{-ab'}{b^2} = \frac{a'b-ab'}{b^2} \]

OpenStudy (anonymous):

Next step is rational exponents.

OpenStudy (anonymous):

We would expect:\[ a^n = a'na^{n-1} \]to still hold true.

OpenStudy (anonymous):

\[ (\sqrt{a})' = a'\frac12a^{-1/2} = \frac{a'}{2\sqrt{a}} \]

OpenStudy (anonymous):

What is weird is how similar it is to derivative, yet anti-derivative has no meaning.

OpenStudy (anonymous):

\[ (a+b)' = a' (1+b/a) + b'(1+a/b) \]

OpenStudy (kainui):

\[a=a^{k/k}=(a^k)^{1/k}\]\[\large a' = \frac{1}{k}(a^k)^{\frac{1}{k}-1}(a^k)'=\frac{1}{k}(a^k)^{\frac{1}{k}-1}*ka^{k-1}a'\] \[\large 1=(a^k)^{\frac{1}{k}-1}a^{k-1}\]\[\large 1=a^{1-k +k-1}\]

OpenStudy (anonymous):

\[ a' + \frac{a'b}{a} + b' + \frac{ab'}{b} = \frac{a'ab + ab^2+abb'+a^2b'}{ab} = (a+b)' \]

OpenStudy (anonymous):

Did I mess this last one up?

OpenStudy (anonymous):

It looks like it could be useful as hell.

OpenStudy (kainui):

Yeah I'm trying to figure it out, I have been wanting that for a while now haha

OpenStudy (anonymous):

What I did was \(a+b = a(1+b/a)=b(1+a/b)\)

OpenStudy (anonymous):

If let \(b=1\), then we get \[ (a+1)' = \frac{a'a+a}{a} \]

OpenStudy (anonymous):

Oh no, did I crack the code?

OpenStudy (anonymous):

I messed up, because it would mean \[ (a+1)' = a' + 1 \]which is not true.

OpenStudy (anonymous):

Oh, I know what I did wrong now.

OpenStudy (anonymous):

I wasn't consistent with my factoring.

OpenStudy (anonymous):

\[ (a+b)' = a'(1+b/a) + a(1+b/a)' \]

OpenStudy (kainui):

I am playing around with the integral and it's sort of bizarre how it doesn't work quite right. \[6'=5=(\frac{1}{2}5^2)'\]

OpenStudy (anonymous):

\[ (a+1)' = a'(1+1/a) + a(1+1/a)' \]

OpenStudy (kainui):

Well I guess it does work, I was thinking as though the 1/2 is a constant... But it's not even a thing here @_@

OpenStudy (anonymous):

\[ a'\left(\frac{1+a}{a}\right) + a\left(\frac{1+a}{a}\right) ' \]

OpenStudy (anonymous):

I think if we can do \((a+1)'\) we win.

OpenStudy (anonymous):

but the \(1+a\) inside the derivative means we'll probably end up at a circular equality.

OpenStudy (kainui):

\[(\frac{1}{2}5^2)' = -(\frac{5}{2})^2+5=-\frac{5}{4} \ne 5\] for the record. Hmm I think I might have a way to make the derivative work. It's just so odd that it's not a linear operator but so many other properties seem to hold pretty well.

OpenStudy (anonymous):

Where did you get \(\frac 12 5^2\)?

OpenStudy (kainui):

In my misguided attempt to integrate 5. I knew 6'=5 and thought I could just increase the exponent by 1 and divide by it.

OpenStudy (kainui):

Another interesting thing, 21'=25'=10

OpenStudy (anonymous):

\[ (a^n)' = a'na^{n-1} \]

OpenStudy (anonymous):

\[ (p^n)' = np^{n-1} \]So you can do the anti derivative if you think it comes from a prime.

OpenStudy (anonymous):

But only \(1\) can come from a prime.

OpenStudy (kainui):

Are there any cases where a'=1 when a is not prime?

OpenStudy (kainui):

Well now that we have negative numbers floating around, maybe it's possible?

OpenStudy (anonymous):

Well, suppose there was a composite \(a\) such that \(a'=1\).

OpenStudy (anonymous):

It would mean that \((a^a)' = a^a\)

OpenStudy (kainui):

Well I'm not just saying composite, but any rational or irrational number we have access to now.

OpenStudy (anonymous):

Oh

OpenStudy (anonymous):

We would need to search for:\[ 1 = \frac{(a^{n})'}{na^{n-1}} \]

OpenStudy (anonymous):

\[ 1=\frac{a'a^{n-1}+a(a^{n-1})'}{na^{n-1}} \]

OpenStudy (kainui):

\[(-1)' = (e^{i \pi})'\]\[0=e' i \pi e^{i \pi -1}\]\[0=e'\]

OpenStudy (anonymous):

If only we figured out: \[ (f(n))' \]

OpenStudy (kainui):

\[i' = (e^{i \pi /2})'\]\[i' = e' i \frac{\pi}{2}e^{i \pi/2 -1}\]\[i'=0\]

OpenStudy (anonymous):

\[ i = (-1)^{1/2} = (-1)'(\ldots) = 0 \]

OpenStudy (kainui):

\[e' = [\sum_{n=0}^\infty \frac{1}{n!} ]'=0\]\[e' = [\lim_{n \rightarrow \infty} (1+\frac{1}{n})^n]'=0\]

OpenStudy (anonymous):

What about \(\pi\) as a series?

OpenStudy (kainui):

I don't know, I think it has several kinds of infinite representations. But I'm more interested in working out some kind of sum rule. I just sorta wanted to throw those out there since they sorta came to mind.

OpenStudy (anonymous):

There is also vectors.

OpenStudy (kainui):

Ohhhh speaking of which: \[\LARGE a+bi = re^{i \theta}\]

OpenStudy (anonymous):

Perhaps gcd thing somehow involves \(aa'\).

OpenStudy (anonymous):

For primes \(aa' = a\).

OpenStudy (kainui):

Interesting thought.

OpenStudy (anonymous):

\[ \langle a',b' \rangle \cdot \langle b,a \rangle = a'b+ab' \]

OpenStudy (anonymous):

We could say \(\mathbf v' \) is just \(\langle v_1',v_2',\ldots\rangle\).

OpenStudy (anonymous):

Unless you think there is a better way of defining it.

OpenStudy (anonymous):

\[ \mathbf v\cdot \mathbf v' = v_1v_1'+v_2v_2'+\ldots \]

OpenStudy (anonymous):

For a vector of primes, it would just be the sum of the elements.

OpenStudy (kainui):

Hmmm well if it is consistent with something that seems natural. I don't know if I'm willing to say anything yet. I'm currently looking at; \[(a+bi)' = (re^{i \theta})'=r'e^{i \theta}+r (e^{i \theta})'=r'e^{i \theta}\] So the derivative seems for complex numbers to be only dependent on the length not the direction.

OpenStudy (kainui):

What I'm saying is the derivative should be the derivative of the length of the vector in the same direction as the old vector.

OpenStudy (anonymous):

And \[ r = \sqrt{a^2+b^2} \]?

OpenStudy (kainui):

Yes, exactly. Or also written as \[r= [(a+bi)(a-bi)]^{1/2}\]

OpenStudy (kainui):

But it follows from this fairly well: \[(re^{i \theta})' = r'e^{i \theta}\] since e'=0

OpenStudy (anonymous):

So you are saying: \[ \mathbf v' =|\mathbf v|' \frac{\mathbf v}{|\mathbf v|} \]

OpenStudy (kainui):

Yeah perfect.

OpenStudy (anonymous):

So when \(|v|\) is prime, you end up with a unit vector.

OpenStudy (kainui):

Yeah exactly so v=<3, 4> has a derivative v' as a unit vector in the same direction

OpenStudy (anonymous):

And the anti derivative of any unit vector is simply to multiply it by any prime number.

OpenStudy (anonymous):

And the derivative of any unit vector is 0 vector.

OpenStudy (kainui):

Oddly enough. So instead of adding +C it's like multiplying by p^1

OpenStudy (anonymous):

I have to wonder now...

OpenStudy (anonymous):

For all \(a\), does there exist an \(n\), such that:\[ a^{(n)} = 1 \]Or do loops exist?

OpenStudy (anonymous):

Well, loops exist for \(p^p\)

OpenStudy (kainui):

Yeah, and some actually increase for instance 12'=16, 16'=32, etc?

OpenStudy (anonymous):

\[ 12=2^23 \\ 16 = 2^4\\ 32 = 2^5 \]

Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!
Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!