Skip to main content
edited tags
Link
GH from MO
  • 105.2k
  • 8
  • 292
  • 398
edited title
Source Link
MrPie
  • 305
  • 1
  • 7

Recently I have made the observationsome interesting observations on the limit $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}. $$ When this limit exist denote itsdenoteits convergence point by $\zeta$. If we define $f_k(n)$$f_k$ to be the partial sums of this series:

$$f_k(n):= \sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}} $$$$f_k(\alpha,\beta):= \sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}} $$ I want to prove that $$f_{2k}(n)\le \zeta \le f_{2k+1}(n) $$$$f_{2k}(\alpha,\beta)\le \zeta \le f_{2k+1}(\alpha,\beta)$$ for all values k. One approach is to prove them separately. First show $\zeta-f_{2k}(n)\ge 0$$\zeta-f_{2k}(\alpha,\beta)\ge 0$ and then $f_{2k+1}(n) - \zeta \le 0$$f_{2k+1}(\alpha,\beta) - \zeta \le 0$.

From this we can easily obtain small bounds on the actual value of the zeta function and hopefully show that the there is only one value of $\alpha$ such that both the limits $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}$$ and $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \sin\left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}$$ are both simultaneously zero. We should be able to bound zeta by the first few terms.

Recently I have made the observation on the limit $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}. $$ When this limit exist denote its convergence point by $\zeta$. If we define $f_k(n)$ to be the partial sums of this series:

$$f_k(n):= \sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}} $$ I want to prove that $$f_{2k}(n)\le \zeta \le f_{2k+1}(n) $$ for all values k. One approach is to prove them separately. First show $\zeta-f_{2k}(n)\ge 0$ and then $f_{2k+1}(n) - \zeta \le 0$.

From this we can easily obtain small bounds on the actual value of the zeta function and hopefully show that the there is only one value of $\alpha$ such that both the limits $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}$$ and $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \sin\left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}$$ are both simultaneously zero. We should be able to bound zeta by the first few terms.

Recently I have made some interesting observations on the limit $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}. $$ When this limit exist denoteits convergence point by $\zeta$. If we define $f_k$ to be the partial sums of this series:

$$f_k(\alpha,\beta):= \sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}} $$ I want to prove that $$f_{2k}(\alpha,\beta)\le \zeta \le f_{2k+1}(\alpha,\beta)$$ for all values k. One approach is to prove them separately. First show $\zeta-f_{2k}(\alpha,\beta)\ge 0$ and then $f_{2k+1}(\alpha,\beta) - \zeta \le 0$.

From this we can easily obtain small bounds on the actual value of the zeta function and hopefully show that the there is only one value of $\alpha$ such that both the limits $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \cos \left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}$$ and $$\lim_{k\rightarrow \infty}{\sum_{n=1}^{k}{\dfrac{(-1)^{n-1}\biggl( \sin\left(\beta\ln(n)\right)\biggr)}{n^{\alpha}}}}$$ are both simultaneously zero. We should be able to bound zeta by the first few terms.

edited title
Link
MrPie
  • 305
  • 1
  • 7

proving inequality in ReimannRiemann zeta function

edited tags
Link
GH from MO
  • 105.2k
  • 8
  • 292
  • 398
Loading
Source Link
MrPie
  • 305
  • 1
  • 7
Loading