0
$\begingroup$

I have the following differential equation $$y''-xy=0\,.$$ By using $y=\sum\limits_{n=0}^{\infty} a_n x^n$, I show that $$(n+2)(n+1)a_{n+2}=a_{n-1}$$ for $n=1,2,...$, and $a_2=0$. (I differentiate the sum and by plugging it into the differential equation which is zero I get that result.) My question is how can I show that solutions converge for all $x \in \mathbb{R} $ since I cannot apply the ratio test.

$\endgroup$
  • $\begingroup$ note that if $a_2=0$ then $a_4=0=a_6=a_8....=0$ $\endgroup$ – Isham Jul 30 '18 at 14:03
  • $\begingroup$ Let $b_n = a_{2n-1}x^{2n-1}$. Since $a_{2n} = 0$, your sum is equivalent to $\sum b_n$ which you can apply the ratio test to. $\endgroup$ – AlexanderJ93 Jul 30 '18 at 14:03
  • $\begingroup$ Is your recurrence relation correct? $\endgroup$ – Nosrati Jul 30 '18 at 14:07
  • 1
    $\begingroup$ Your recurrence seems not right…? 'Cause for $y''$, the coefficient of $x^n$ is $(n+2) (n+1) a_{n+2}$; for $xy$, the corresponding coefficient is $a_{n-1}$, so… $\endgroup$ – xbh Jul 30 '18 at 14:08
  • $\begingroup$ For ratio test, maybe you could solve the recurrence first then test it? $\endgroup$ – xbh Jul 30 '18 at 14:10
1
$\begingroup$

There are two linearly independent solutions $f_0$ and $f_1$. We assume that the first one $f_0(x)$ satisfies the conditions $f_0(0)=1$ and $f_0'(0)=0$, whilst the second one $f_1(x)$ satisfies the conditions $f_1(0)=0$ and $f_1'(0)=1$. Any solution $y=f(x)$ of this differential equation is written as the sum $$f(x)=f(0)\,f_0(x)+f'(0)\,f_1(x)\text{ for all }x\in\mathbb{R}\,.$$ If we show that $f_0$ and $f_1$ are analytic on the whole $\mathbb{R}$, then we are done.

Now, from the result you obtain regarding the relations of the coefficients, you have that $$f_0(x)=\sum_{k=0}^\infty\,u_{k}\,x^{3k}\text{ and }f_1(x)=\sum_{k=0}^\infty\,v_{k}\,x^{3k+1}\,,$$ where $$u_k:=\frac{1}{\prod\limits_{r=1}^k\,\big((3k)(3k-1)\big)}=\frac{1}{(3k)_{k,3}\,(3k-1)_{k,3}}$$ and $$v_k:=\frac{1}{\prod\limits_{r=1}^k\,\big((3k+1)(3k)\big)}=\frac{1}{(3k+1)_{k,3}\,(3k)_{k,3}}\,.$$ Here, the falling factorial power $(x)_{n,h}$ for $x,h\in\mathbb{C}$ and $n\in\mathbb{Z}_{\geq 0}$ is defined by $$(x)_{n,h}:=x\,(x-h)\,(x-2h)\,\ldots\,\big(x-(n-1)h\big)\,,$$ with $(x)_{0,h}:=1$.

To show that the radii of convergence of $f_0$ and $f_1$ are both infinite, we need to verify that $$\limsup_{k\to\infty}\,\sqrt[3k]{u_k}=0\text{ and }\limsup_{k\to\infty}\,\sqrt[3k+1]{v_k}=0\,.$$ However, this follows from the observation that $$(3k+1)_{k,3}\geq k!\,,\,\,(3k)_{k,3}\geq k!\,,\text{ and }(3k-1)_{k,3}\geq k!$$ for all $k=0,1,2,\ldots$, and the well known fact that $$\sqrt[k]{k!}\to\infty\text{ as }k\to\infty\,.$$


It turns out that $$f_0(x)=\frac{\sqrt[3]{3}^2\,\Gamma\left(\frac{2}{3}\right)}{2}\,\text{Ai}(x)+\frac{\sqrt[6]{3}\,\Gamma\left(\frac{2}{3}\right)}{2}\,\text{Bi}(x)$$ and $$f_1(x)=-\frac{\sqrt[3]{3}\,\Gamma\left(\frac{1}{3}\right)}{2}\,\text{Ai}(x)+\frac{\Gamma\left(\frac{1}{3}\right)}{2\,\sqrt[6]{3}}\,\text{Bi}(x)\,,$$ where $\text{Ai}$ and $\text{Bi}$ are Airy functions, and $\Gamma$ the usual gamma function. Alternatively, $$\text{Ai}(x)=\frac{1}{\sqrt[3]{3}^2\,\Gamma\left(\frac23\right)}\,f_0(x)-\frac{1}{\sqrt[3]{3}\,\Gamma\left(\frac13\right)}\,f_1(x)$$ and $$\text{Bi}(x)=\frac{1}{\sqrt[6]{3}\,\Gamma\left(\frac23\right)}\,f_0(x)+\frac{\sqrt[6]{3}}{\Gamma\left(\frac13\right)}\,f_1(x)\,.$$

$\endgroup$

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.