1

Define $\gamma=\lim_{n\to\infty} \sum_{k=1}^n 1/k - ln(n)$.

I know that $\gamma$ is nonnegative, but i don't know how to prove that it is positive.

John. p
  • 1,663

1 Answers1

0

A simple comparison suffices: consider the functions $$f(x) = \frac{1}{x}, \quad g(x) = \frac{1}{\lfloor x \rfloor}.$$ Clearly, $g(x) \ge f(x)$ for all $x \ge 1$. Integrating both on $[1, n]$ gives $$\int_{x=1}^n f(x) \, dx \le \int_{x=1}^n g(x) \, dx,$$ or $$\log n \le \sum_{k=1}^{n-1} \frac{1}{k} < \sum_{k=1}^n \frac{1}{k}.$$

heropup
  • 135,869
  • I know that $\sum_{k=1}^n \frac{1}{k} - \log n >1/n$. But how does that gurantee that its limit is positive – John. p Mar 30 '14 at 04:46
  • Think about the picture: Consider $$D_k = \int_{x=k}^{k+1} \frac{1}{\lfloor x \rfloor} - \frac{1}{x} , dx = \frac{1}{k} - \log(k+1) + \log k > 0.$$ Then the limit of the partial sums of $S_n = \sum_{k=1}^n \frac{1}{k} - \log (n+1)$ is simply $$S_\infty = \sum_{k=1}^\infty D_k > 0,$$ and your sum is even larger than that. – heropup Mar 30 '14 at 04:51
  • see http://math.stackexchange.com/questions/306371/simple-proof-of-showing-the-harmonic-number-h-n-theta-log-n/306379#306379 – Will Jagy Mar 30 '14 at 04:52