Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

If we build the sum of two power series , $\sum a_kz^k$, $\sum b_kz^k$ both with radius of convergence 1, why does the resulting power series: $$\sum (a_k+b_k)z^k$$ also have RoC 1?

Is it because : $$|\frac{a_k}{a_{k+1}}| = 1 , |\frac{b_k}{b_{k+1}}|=1$$

$$|\frac{a_k+b_k}{a_{k+1}+b_{k+1}}|$$

that doesnt really do anything T_T .... If you see how to show this, please do tell.

share|improve this question
    
Not true. The series has radius of convergence at least 1, but it may very well be larger. –  Andres Caicedo Sep 13 '12 at 0:18
add comment

1 Answer 1

up vote 3 down vote accepted

The radius of convergence need not stay the same.

Let $p(z)$ have radius of convergence $1$, and let $q(z)$ have radius of convergence $4$. Then $(q+p)(z)$ and $(q-p)(z)$ have radius of convergence $1$, but their sum has radius of convergence $4$.

For any $k\ge 1$, including "$k=\infty$", we can find power series with radius of convergence $1$ whose sum has radius of convergence $k$.

share|improve this answer
    
I dont see, why cant the sum be smaller than one? And how would you create one with k= infinity if both are required to have the same radius of convergence (by setting $b_k = -a_k$ ? ) –  bakabakabaka Sep 13 '12 at 0:32
    
@bakabakabaka: Sure, $b_k=-a_k$ is good. For for a fancier example, $e^z-\frac{1}{1-z}$ and $e^z+\frac{1}{1-z}$. As for your question about why not less than $1$, the sum has radius of convergence $\ge 1$ by say the ratio test. –  André Nicolas Sep 13 '12 at 0:39
    
How does one find the radius of convergence of the sum ? F.e. how did you know that q+p had radius of convergence 1 ? 1+4 would be 5, no? Same with q-p , too, it would be 3 ? –  bakabakabaka Sep 13 '12 at 1:07
    
@bakabakabaka: Formally, use Ratio Test or Root Test. More informally, if $p(z)$ is nice only when $|z|\lt 1$, and $q(z)$ is nice for $|z|\lt 4$, then $q(z)+p(z)$ is nice only when $|z|\lt 1$. To put it another way, $p(z)$ has a singularity somewhere on the circle $|z|=1$, and $q(z)$ is OK for any $z$ with $|z|\lt 4$, so $q(z)+p(z)$ goes bad somewhere on $|z|=1$. –  André Nicolas Sep 13 '12 at 1:17
1  
@bakabakabaka: Informal argument for $q-p$ is the same. For formal, Ratio Test on coefficients is unwieldy, so here is another way. Let $\sum p_n z^n$ have radius of convergence $1$, let $\sum q_n z^n$ have radius of convergence $4$. So for example $\sum p_n z^n$ diverges if $|z|=1.01$. But $\sum q_n z^n$ converges for $|z|=1.01$. It follows that $\sum (p_n+q_n)z^n$ diverges for $|z|=1.01$. So radius of convergence of sum is $\le 1.01$. In same way, it is $\le 1+\epsilon$ for every positive $\epsilon$, so it is $\le 1$. –  André Nicolas Sep 13 '12 at 1:54
show 2 more comments

Your Answer

 
discard

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.