Linearly Independent of exponenial functions

Show 40 post(s) from this thread on one page
Page 1 of 2 12 Last
• Nov 30th 2010, 04:55 AM
Liverpool
Linearly Independent of exponenial functions
The problem :

Show that the functions $e^{ax}$, $e^{bx}$ and $e^{cx}$ are linearly independent if $a \neq b \neq c$

Solution :
I calculated the wronskian, and my final answer is:

$e^{(a+b+c)x} \left( (b-a)c^2+(a-c)b^2+(c-b)a^2 \right)$

Clearly, $e^{(a+b+c)x} \neq 0$

The problem here is with $(b-a)c^2+(a-c)b^2+(c-b)a^2$

Since $a \neq b \neq c$, the following numbers can't be zero:
• b-a
• a-c
• c-b
Now, I stopped!

How I can prove that the expression $(b-a)c^2+(a-c)b^2+(c-b)a^2$ can't be zero so that wronskian will be non zero and hence the function are linearly independent?
• Nov 30th 2010, 05:01 AM
Ackbeet
[EDIT]: Your Wronskian is correct, just not simplified as much as it could be. If you simplify it a bit further, I think your result will pop right out.
• Nov 30th 2010, 05:07 AM
Liverpool
I do not see any further simplification :S
• Nov 30th 2010, 05:33 AM
Ackbeet
Well, you're trying to set the expression equal to zero, and see under what conditions that can happen, right? So you have a quadratic in $a$, for example. Try solving

$(b-a)c^{2}+(a-c)b^{2}+(c-b)a^{2}=0$ for $a.$

What do you get?
• Nov 30th 2010, 09:41 AM
Liverpool
it will be:

$(c-b)a^2+(b^2-c^2)a+(bc^2-cb^2)=0$

here:

$a'=c-b , b'=b^2-c^2 & c'=bc^2-cb^2$

so:

$a_{1,2}=\dfrac{-b' \pm \sqrt{ (b')^2 - 4a'c' } } { 2a' }$

it will be too long to solve it since i should find $(b')^2 , 4a'c'$ ... etc

Anyone could explain?
• Nov 30th 2010, 09:51 AM
Ackbeet
You should carry your computation through. It's not that much algebra. I think you'll find that your answers simplify down a great deal.
• Nov 30th 2010, 09:53 AM
Liverpool
This is what I can do.
I do not see any further simplification.
• Nov 30th 2010, 09:56 AM
Ackbeet
Just plug in your primed variables. Try things. Note that $a^{2}-b^{2}=(a-b)(a+b).$

The simplification is, for me, about 5 lines' worth. If you're taking differential equations, you should definitely be prepared to do that much algebra.

This forum is about helping people get unstuck. It is not about giving people answers. So plug things in, show your work, and let me know if you get stuck.
• Nov 30th 2010, 10:05 AM
Liverpool
xD.
yup yup it will be reduced to:

$a^2-(b+c)a+bc=0$

So the solution for this equation is:

$a_{1,2}=\dfrac{b+c \pm (b-c)}{2}$

so ?
• Nov 30th 2010, 10:08 AM
Ackbeet
Keep going. Take the plus sign solution: what is a? Take the negative sign solution: what is a?
• Nov 30th 2010, 10:09 AM
Liverpool
The two solution are: b and c.
so?
• Nov 30th 2010, 10:11 AM
Ackbeet
So, you could factor the original expression

$(b-a)c^{2}+(a-c)b^{2}+(c-b)a^{2}$

as

$(a-c)(a-b)x,$

where $x$ is something or other. Do you think you could figure out what $x$ is? What do you want it to be?
• Nov 30th 2010, 10:14 AM
Liverpool
Well,

$x=\dfrac{(b-a)c^{2}+(a-c)b^{2}+(c-b)a^{2}}{(a-c)(a-b)}$

I do not what I want it to be!
• Nov 30th 2010, 10:18 AM
Ackbeet
Consider the symmetry of the original expression. a, b, and c are all treated the same, right? You could probably replace a with b, b with c, and c with a without changing the expression (or maybe you'd get a minus sign - that's not important). But (a-c)(a-b) is not that symmetric. How could you make it equally symmetric? Another way of phrasing it is this: what is missing?
• Nov 30th 2010, 10:21 AM
Liverpool
Sorry, but I did not understand "Consider the symmetry of the original expression".
Show 40 post(s) from this thread on one page
Page 1 of 2 12 Last