1
$\begingroup$

Solving this differential equation with an online calculator:

$$-(a z+b) y+(c z+d) y''+cy' = 0$$

I obtain something like:

$$y(z)=C_1 \exp\left(\frac{-\sqrt{a}z}{\sqrt{c}}\right) U(arg1,arg2,arg3)+C_2 \exp\left(\frac{-\sqrt{a}z}{\sqrt{c}}\right) L_{arg1}(arg2)$$

with arg are arguments of common functions.

I have two problems:

  • I don't know the functions U and L
  • if $c<0$ is it possible to find a real solution for $y(z)$ playing with the $C_1$ and $C_2$ parameters ?
$\endgroup$
  • $\begingroup$ $U$ is the confluent hypergeometric function ; $L$ is the generalized Laguerre polynomial $\endgroup$ – Claude Leibovici Nov 25 '14 at 11:09
0
$\begingroup$

It seems that you were using wolframalpha. There are links about the functions $U$ and $L$ - for example, Confluent Hypergeometric Function of the Second Kind and Laguerre Polynomial .

In the case where $c=0$ we have to work with Airy functions, which are, essentially, the independent solutions of the the equation $y''(z)=zy$.

As for the existence of real solutions, consider the Cauchy problem for your equation with initial data $y(z_0)=y_0$. If $cz_0+d\ne 0$, then your problem is well-posed and therefore by Cauchy-Lipschitz-Lindelof theorem has local solution, which would be real. You can build a maximum real solution afterwards. I don't know if this maximum solution will be defined on an interval, half-line or the whole line.

$\endgroup$
  • $\begingroup$ I already solved the case $c = 0$ with the Airy functions. The next step was indeed to work with the following conditions: -$c \ne 0$ -$y(z=0) = y_0$ -$y^{'}(z=z_0)=0$ I also have $c<0$ $a>0$ b and d are non-zero parameters. How to proceed ? The $C_1$ and $C_2$ constant may be foud analytically ? $\endgroup$ – user3473016 Nov 25 '14 at 11:20
  • $\begingroup$ And what if the argument 1 of the Laguerre polynomial is not an integer, which is the case in my situation ? I only use the first part of the solution ? $\endgroup$ – user3473016 Nov 25 '14 at 13:16
  • $\begingroup$ @user3473016 this, unfortunately, I don't know. There must be some generalisations of Laguerre polynomials that accept non-integer parameters. $\endgroup$ – TZakrevskiy Nov 25 '14 at 13:26

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.