1
$\begingroup$

The question is as above. I have not learnt about the Wronskian or the likes.

I suppose that I'll have to do something with differentiation or trying different x values.

EDIT:

  • $a$ and $b$ are non-zero and not equal to each other. I should have mentioned this earlier.

  • I have not learnt about the Wronskian formally and doubt that I should use it. It is nice to learn new methods and new theorems, however.

$\endgroup$
2
  • 1
    $\begingroup$ I assume that $a \neq b$ and both are non-zero? In that case suppose they are linearly dependent and show a contradiction. $\endgroup$
    – copper.hat
    Apr 16, 2015 at 0:30
  • $\begingroup$ A further hint: If they are linearly dependent, then there exists $c_1, c_2, c_3$ (with at least one of $c_i$ nonzero) such that $c_1 5 + c_2 e^{ax}+c_3e^{bx}=0$. $\endgroup$
    – JMoravitz
    Apr 16, 2015 at 0:40

3 Answers 3

Reset to default
2
$\begingroup$

Let they be linearly dependent, i.e. $$ A\cdot5+Be^{ax}+Ce^{bx}\equiv0 $$ for $A$, $B$, $C$ not all equal to zero. Then it is true for example for $x=0$, 1, 2, i.e. $$ 5A+B+C=0, 5A+Be^a+Ce^b=0, 5A+Be^{2a}+Ce^{2b}=0. $$ The determinant is equal to $$ \begin{vmatrix} 5 & 1 &1\\ 5 & e^a &e^b\\ 5 & e^{2a} & e^{2b} \end{vmatrix}. $$ One can verify (please compute) that it in nonzero. But it means, that there is only one solution $A=B=C=0$. A contradiction.

BTW: Methods with wronskian are easier to use.

$\endgroup$
2
$\begingroup$

(I'm assuming $a\ne b$, since otherwise they are dependent.)

One thing you might notice is that these all grow at different rates. A way to use that is the following:

If they are linearly dependent, we would have: $$A\cdot5+Be^{at}+Ce^{bt}=0$$ where the coefficients are nonzero.

WLOG, assume $a<b$. (If $a>b$, just switch around $a$ and $b$ in the following argument.) I'll also assume that they are both positive, but it's not hard to modify the argument for if they're negative.

Divide everything by $e^{bt}$ (we can do this since it's never equal to zero): $$5Ae^{-bt}+Be^{(a-b)t}+C=0$$ Taking the limit as $t\to\infty$, we have: $$C=0$$ which contradicts the coefficients being nonzero!

$\endgroup$
2
  • $\begingroup$ where the coefficients are nonzero -> isn't this wrong? Linear dependence would mean that not all are zero, some may be. $\endgroup$
    – agdhruv
    Mar 25, 2018 at 20:34
  • $\begingroup$ You are absolutely correct. I was mistaken. $\endgroup$
    – Akiva Weinberger
    Mar 25, 2018 at 20:38
0
$\begingroup$

If any pair of these are linearly independent, then $cy_1 \neq y_2$, for any $c$. So, for example, $ce^{ax}\neq e^{bx}$. Suppose the opposite, $ce^{ax}=e^{bx}$. Dividing by $e^{ax}$, $c=e^{(b-a)x}$. But $c$ is a constant, it can't be a function of $x$. So we have a contradiction (if $a\neq b$), so they are linearly indpendent, and it works the same way if you use $5$ instead of one of the exponential functions.

$\endgroup$
1
  • $\begingroup$ This proves that they are pairwise linearly independent, but doesn't yet prove they are mutually linearly independent. It requires a bit more effort to show that the three together can't combine somehow to equal zero. $\endgroup$
    – JMoravitz
    Apr 16, 2015 at 0:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.