Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

use power series method to find solution of $(x^2+2x)y''+(x^2-2)y'-(2x+2)y=0$ i have tried $y=\sum_{0}^{\infty}{am*x^{m+r}}$ normalize the power to $m+r-1$ ,for m=0, $(2r^2-4r)a_0=0$, $r=0,2$ or a0=0 , m=1 , $(r^2-r+2)a_0=0$



then i don't know the next step to find the regular pattern, can anyone help me?

share|cite|improve this question
I have no idea how you get the identity just after "$m\ge2$". Can you explain? – Did Dec 6 '13 at 7:11
sorry,i have edited the formula again. – user113644 Dec 6 '13 at 8:09

The approach suggested to you, only implemented more rigorously, is the following. Assume that $y(x)=\sum\limits_na_nx^{n+r}$ solves the ODE, then $$y'(x)=\sum\limits_n(n+r+1)a_{n+1}x^{n+r},\quad y''(x)=\sum\limits_n(n+r+2)(n+r+1)a_{n+2}x^{n+r}, $$ and, for every integer $i$, $$ x^iy(x)=\sum\limits_na_{n-i}x^{n+r},\quad x^iy'(x)=\sum\limits_n(n+r+1-i)a_{n+1-i}x^{n+r}, $$ and $$ x^iy''(x)=\sum\limits_n(n+r+2-i)(n+r+1-i)a_{n+2-i}x^{n+r}. $$ Hence, the first term of the ODE becomes $$ (x^2+2x)y''(x)=\sum\limits_n(n+r)(n+r-1)a_{n}+2(n+r+1)(n+r)a_{n+1} x^{n+r}, $$ and similarly for the other terms. Putting everything together yields a relation between $a_n$, $a_{n+1}$ and $a_{n-1}$, to be solved. The initial condition comes from the terms of lowest degree. These are the initial term of $2xy''-2y'$, thus $$ 2r(r-1)a_0-2ra_0=0, $$ which yields $r=0$ or $r=2$, hence everything is in place to solve the ODE.

Another approach, more direct, is to note that, considering $a(x)=x^2+2x$, the ODE becomes $$ a(x)y''(x)+(a(x)-a'(x))y'(x)-a'(x)y(x)=0, $$ that is, considering $z(x)=y'(x)+y(x)$, $$ a(x)z'(x)-a'(x)z(x)=0. $$ Thus, there exists some constant $\lambda$ such that $$z(x)=\lambda a(x). $$ Now, $a(x)=b(x)+b'(x)$ where $b:x\mapsto x^2$ hence there exists some $\lambda$ such that $$ y'(x)+y(x)=\lambda (b(x)+b'(x)), $$ that is, $$ (y-\lambda b)'(x)=-(y-\lambda b)(x), $$ hence there exists some $\mu$ such that $$ (y-\lambda b)(x)=\mu\mathrm e^{-x}, $$ that is, $$ y(x)=\lambda x^2+\mu\mathrm e^{-x}. $$ Note that the case $r=0$ in the first approach corresponds to $\mu\ne0$ while the case $r=2$ corresponds to $\mu=0$ and $\lambda\ne0$.

share|cite|improve this answer
thank you so much!!the second method is so amazing! – user113644 Dec 7 '13 at 1:59

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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