Is there a simple way to express this? \[\Large (A+B)^k\]
\[\Large (A+B)^k\] A and B are matrices and k is positive integer. We can't assume that A and B commute.
how can A+B not commute
They're matrices...
matrices commute under addition
\[\Large (A+B)^2 \ne A^2+2AB+B^2\]
Oh i thought u meant it cant be written (B+A)^k
how about diagonlizing both
Hmm well that's a good idea but technically I'm not really interested in matrices I'm interested in noncommuting operators and thought this question would be a simple analogy to see if we could figure out a binomial theorem to use instead. So can we do it without diagonalizing them, but still have some kind of binomial theorem form? haha
what happens when you expand ( A + B ) * ( A + B ) using matrix multiplication rules distributivity
is that equal to (A+B) * A + (A+B)*B A*A + B*A + A*B + B*B ?
Ok so we can conclude this Can ::) assune they are diagonalizable?
u can do that perl
My real question is finding the general form of this derivative, I expressed it as this \[\Large \dfrac{d^k}{dx^k}\left( e^{f(x)} \right)=e^{f(x)}(f'(x)\cdot+\frac{d}{dx}\cdot)^k\] So when k=1 we have: \[\large e^{f(x)} (f'(x)+\frac{d}{dx})1=e^{f(x)} (f'(x)*1+\frac{d}{dx}1)=e^{f(x)}f'(x)\]
how to write a general form for (A+B)^k is what we gotta find
Yeah you are right @perl on distributing it out, it is A^2+AB+BA+B^2
ok i see , so we can't assume A, and B commute under multiplication, so it doesn't simplify to A^2 + 2AB + B^2
For comparison, \[e^{f(x)} (f'(x)+\frac{d}{dx})(f'(x)+\frac{d}{dx})1 = e^{f(x)}(f'(x)^2+f''(x)+f'(x)\frac{d}{dx}+\frac{d^2}{dx^2}) \\ e^{f(x)}(f'(x)^2+f''(x))=\frac{d^2}{dx^2}(e^{f(x)})\]
\(e^A=\phi(A)\)
So yeah, the A and B thing is really a lot simpler than all this nonsense, unless you can figure out how to diagonalize either of these, multiplication by f'(x) or the derivative operator.
such that \(\phi(A)\) is a fundamental solution matrices
Hmm well here f(x) is just a regular function of x, not a matrix at all.
hmm :\
Really, the general form of this derivative is what I'm after where f(x) is a function. \[\Large \frac{d^k}{dx^k}(e^{f(x)})\] So if you can help me solve this or the matrix problem would be helping me the same haha.
I think repeated logarithmic differentiation would work pretty well here.
Oh that sounds promising, can you explain what you mean by that?
Do you know what logarithmic differentiation is?
Yes, but I'm not sure how I would apply it to this situation.
Oh another fact, we can assume I already know all the derivatives of f(x), meaning f'(x), f''(x), f'''(x), etc...
Nothing special. Just brute force some logarithmic derivative for a few times and try to find a closed form solution lol.
Yeah brute force how, I'm not really seeing that lol.
the only way i see is expanding :|
Oh like a power series? I'll try that
and we can state special cases:- (it simplify it very good) A,B diagonalizable A,B upper triangle (or the other type) A,B diagonal
i'll take snap for my work
Hmm well I don't really have a clue how we could say multiplication by f'(x) is a diagonal matrix or that the derivative of a function is a upper triangular matrix or something like that... Unless maybe we somehow use a polynomial with an infinite matrix...
hmm no it work with n*n hmm but i see f(x) and f'(x) as vectors not scalar
By the way, the function f(x) has its derivatives defined as this: \[\Large f^{(k)}(x) = \frac{a}{(x+b)^k}\] for k>0
too much work -.-
Wait show me what you have @Marki it might be helpful since I figured out a way to make an upper triangular matrix out of the derivative operator
i just dumbly expanded it
For reference: \[ \begin{align*} \frac{d}{dx}e^{g(x)}=&\,\,e^{g(x)} g'(x)\\ \frac{d^2}{dx^2}e^{g(x)}=&\,\,e^{g(x)} g''(x)+e^{g(x)} g'(x)^2\\ \frac{d^3}{dx^3}e^{g(x)}=&\,\,e^{g(x)} g^{(3)}(x)+e^{g(x)} g'(x)^3+3 e^{g(x)} g'(x) g''(x)\\ \frac{d^4}{dx^4}e^{g(x)}=&\,\,e^{g(x)} g^{(4)}(x)+3 e^{g(x)} g''(x)^2+e^{g(x)} g'(x)^4+4 e^{g(x)} g^{(3)}(x) g'(x)\\&+6 e^{g(x)} g'(x)^2 g''(x)\\ \frac{d^5}{dx^5}e^{g(x)}=&\,\,e^{g(x)} g^{(5)}(x)+e^{g(x)} g'(x)^5+5 e^{g(x)} g^{(4)}(x) g'(x)\\ &+10 e^{g(x)} g^{(3)}(x) g''(x)+10 e^{g(x)} g^{(3)}(x) g'(x)^2+10 e^{g(x)} g'(x)^3 g''(x)\\ &+15 e^{g(x)} g'(x) g''(x)^2\\ \frac{d^6}{dx^6}e^{g(x)}=&\,\,e^{g(x)} g^{(6)}(x)+10 e^{g(x)} g^{(3)}(x)^2+15 e^{g(x)} g''(x)^3+e^{g(x)} g'(x)^6\\ &+6 e^{g(x)} g^{(5)}(x) g'(x)+15 e^{g(x)} g^{(4)}(x) g''(x)+15 e^{g(x)} g^{(4)}(x) g'(x)^2\\ &+20 e^{g(x)} g^{(3)}(x) g'(x)^3+15 e^{g(x)} g'(x)^4 g''(x)+45 e^{g(x)} g'(x)^2 g''(x)^2\\ &+60 e^{g(x)} g^{(3)}(x) g'(x) g''(x) \end{align*} \]
\[\Large \left[\begin{matrix}0 & 1 &0 & 0 \\ 0 & 0 &2 & 0 \\ 0 & 0 &0 & 3 \\ 0 & 0 &0 & 0 \end{matrix}\right] \left[\begin{matrix}1 \\x \\ x^2 \\ x^3\end{matrix}\right]=\left[\begin{matrix}1 \\x \\ x^2 \\ x^3\end{matrix}\right]\] Ok I don't really know how to make the notation correct to saywhat I mean but I have seen this before but never really used it. I think I need to switch these out with the basis and put the coefficients in here, but that's kind of difficult
omg the latex is high
for triangle (A+B)^k |dw:1422785712510:dw|
typo |dw:1422785945094:dw|
Join our real-time social learning platform and learn together with your friends!