0

Theorem: convergence for the ratio test implies convergence for the root test. So whenever the ratio test works (i.e. tells you whether the series converges), the root test also works and the limits coincide.

If $$\lim_{n\to\infty}\frac{|a_{n+1}|}{|a_n|}=L,$$ then $$\lim_{n\to\infty}|a_n|^{1/n}=L.$$

Question: a series that cannot be evaluated through the root test because algebra won't allow it (usually involving a factorial), but can be evaluated through the ratio test will it theoretically have the same limit value through root test (if it were possible to be evaluated) as ratio test due to the theorem?

Example: $$\displaystyle\sum\frac{n!}{n^n}$$

ratio test yields $ L = \frac{1}{e} < 1$

but if you tried the root test, you would be miserable. The limit is 1/e by ratio test implies root test?

(though you can just use Stirling's)

$$\frac{\sqrt{2\pi n} e^{-n} n^n}{n^n} \\ = \sqrt{2\pi} \cdot \frac{1}{e^n} \cdot \sqrt{n}$$

The limit $1 \le \sqrt{n}^{1/n} \le n^{1/n} \to 1$ as $n \to +\infty$ allows us to recover the ratio $1/e$

$$L = \lim\limits_{n\to+\infty} a_n^{1/n} = \lim\limits_{n\to+\infty} \frac1e \sqrt{n}^{1/n} = \frac1e$$

user29418
  • 1,087
  • If I understand you correctly, in the first part of your post you claim that you know that if A then B. Then, in the second part, you ask us whether it is true that when we have A then we also have B. Am I right? – José Carlos Santos May 19 '20 at 10:24
  • yes, but in the beginning, I am assuming both can be computed out analytically but I wanted to check what if it was not computable via root test. – user29418 May 19 '20 at 10:26
  • 1
    Hint : The fact that ration test implies root test is just Cesaro's theorem. Find a counterexample of Cesaro's reciprocal and you will get a counterexample to the root-to-ratio implication. – TheSilverDoe Oct 22 '20 at 16:13

1 Answers1

0

The ratio test can fail but the root test can succeed making the root test stronger than the ratio test.

$ \sum_{n=1}^{\infty} 2^{(-1)^n-n}$

user29418
  • 1,087