My Math Forum I have some simple questions about the Riemann Hypothesis
 User Name Remember Me? Password

 Complex Analysis Complex Analysis Math Forum

 August 10th, 2016, 12:19 PM #1 Senior Member   Joined: Mar 2015 From: England Posts: 163 Thanks: 4 I have some simple questions about the Riemann Hypothesis I have watched and read about Riemann's Zeta function, but in the weeks I've been trying to learn it there are some things I can't quite get my head around. Two things to be exact. One, The Trivial and non trivial Zeta functions that equal to 0. Do they converge to 0? Two, what does this gap on the vertical line mean on Graphs? | | | | / / | | | | After first seeing Riemann's translated original paper I was shocked to see that he didn't have a visual graph of his function.
 August 11th, 2016, 06:03 AM #2 Senior Member   Joined: Mar 2015 From: England Posts: 163 Thanks: 4 Thinking about the trivial zeroes with complex numbers (-2+i0) I look at 2/(2^-2) = 0.25 and I think oh ok so it's going down to zero. But then I look at 1 / 2^-2 and this equals 4. Is the 1 at the top just supposed to represent that it's a fraction of one quarter (1/4)? EDIT: I guess what I'm asking is, how do the trivial and non trivial zeros become zero? Last edited by HawkI; August 11th, 2016 at 06:10 AM.
 September 1st, 2016, 11:57 AM #3 Math Team     Joined: Aug 2012 From: Sana'a , Yemen Posts: 1,177 Thanks: 44 Math Focus: Theory of analytic functions In order to see the trivial zeros of the zeta function. Look at the functional equation $$\zeta(s)=2(2\pi)^{s-1}\Gamma(1-s)\sin(\tfrac12\pi s)\zeta(1-s).$$ You can see that $\zeta (-2n) =0$.
 September 10th, 2016, 11:48 PM #4 Senior Member     Joined: Apr 2014 From: zagreb, croatia Posts: 227 Thanks: 30 Math Focus: philosophy/found of math, metamath, logic, set/category/order/number theory, algebra, topology If I'm not mistaken, (since the series 1 + 1/2^s + 1/3^s... converges for s > 1), and this can be proven as follows $\displaystyle \int$ a to inf dx/x^p, a, p > 0 and not 1 $\displaystyle \int$ a to B dx/x^p = ( 1/(1-p) 1/(x^(p-1))) taken at B minus taken at a (Leibniz-Newton formula) = 1/(1-p) (B^(1-p) - a^(1-p)) If 1-p < 0, the integral converges to 1/(p-1) a^(1-p) because B^(1-p) -> 0 when x -> inf. Now all that remains is to prove Cauchy's integral test for series convergence Let f:I->[0, inf) be a continuous function on I = [1, inf) and let it decrease on I. $\displaystyle \sum f (n)$ n € N converges iff $\displaystyle \int$ 1 to inf f(x) dx converges. Proof. Let s_n = f(1)+...+f(n) be nth partial sum of the series and let F(x) = $\displaystyle \int$ 1 to x f(t) dt, x € [1, inf) Since f decreases f(k+1) <= f(t) <= f(k) for all k € N and all t € [k, k+1], so f(k+1) <= $\displaystyle \int$ k to k+1 f(t) dt <= f(k), k € N So F(n+1) = $\displaystyle \int$ 1 to n+1 f = $\displaystyle \int$ 1 to 2 f +...+ $\displaystyle \int$ n to n+1 f <= f(1) + ...+ f(n) = s_n <= f(1) + $\displaystyle \int$ 1 to 2 f +...+ $\displaystyle \int$ n-1 to n f = = f(1) + F(n) So F(n+1) <= s_n <= f(1) + F(n). If the integral converges, since F grows, it follows from the last inequalities s_n <= f(1) + F(n) <= f(1) + lim n tends to inf F(n) = = f(1) + $\displaystyle \int$ 1 to inf f(x) dx, which shows that the sequence s_n in bounded from above. Since it grows, it's convergent, so the inequalities for n->inf imply $\displaystyle \int$ 1 to inf f(x) dx <= $\displaystyle \sum f (n)$ <= f(1) + $\displaystyle \int$ 1 to inf f(x) dx. So, the series converges. So we get a set that has a point (I don't know the English term, I'll call it limiting point) such that for its any neighbourhood there's an element of it. (Any point s => 1). By a theorem from complex analysis, which states: Let f and g be analytic functions on a domain O (in this case it means an unempty open subset of C such that for any 2 points a, b in it there are finitely many points a = z0, z1, ..., zn = b such that [z0, z1], ...[zn-1, zn] lie in O) If f, g are equal on an infinite set which has a limiting point in O, then theyre equal on O, i.e. f =g. So this means since the series is an analytic function of s > 1) if we extend it to C\{1} that whichever function equals it for s > 1, they're the same. Or, in other words, we can uniquely perform analytic continuation, if we can do it at all. (Which we can and is exectly what Riemann did first I think in the legendary paper you mentioned). And since every analytic function is continuous, by a theorem which states that any value of a continuous function equals its limit at that point, the zeros equal the limit of the function at them.
September 11th, 2016, 01:32 AM   #5
Senior Member

Joined: Apr 2014
From: zagreb, croatia

Posts: 227
Thanks: 30

Math Focus: philosophy/found of math, metamath, logic, set/category/order/number theory, algebra, topology
Quote:
 Originally Posted by raul21 If I'm not mistaken, (since the series 1 + 1/2^s + 1/3^s... converges for s > 1), and this can be proven as follows $\displaystyle \int$ a to inf dx/x^p, a, p > 0 and not 1 $\displaystyle \int$ a to B dx/x^p = ( 1/(1-p) 1/(x^(p-1))) taken at B minus taken at a (Leibniz-Newton formula) = 1/(1-p) (B^(1-p) - a^(1-p)) If 1-p < 0, the integral converges to 1/(p-1) a^(1-p) because B^(1-p) -> 0 when x -> inf. Now all that remains is to prove Cauchy's integral test for series convergence Let f:I->[0, inf) be a continuous function on I = [1, inf) and let it decrease on I. $\displaystyle \sum f (n)$ n € N converges iff $\displaystyle \int$ 1 to inf f(x) dx converges. Proof. Let s_n = f(1)+...+f(n) be nth partial sum of the series and let F(x) = $\displaystyle \int$ 1 to x f(t) dt, x € [1, inf) Since f decreases f(k+1) <= f(t) <= f(k) for all k € N and all t € [k, k+1], so f(k+1) <= $\displaystyle \int$ k to k+1 f(t) dt <= f(k), k € N So F(n+1) = $\displaystyle \int$ 1 to n+1 f = $\displaystyle \int$ 1 to 2 f +...+ $\displaystyle \int$ n to n+1 f <= f(1) + ...+ f(n) = s_n <= f(1) + $\displaystyle \int$ 1 to 2 f +...+ $\displaystyle \int$ n-1 to n f = = f(1) + F(n) So F(n+1) <= s_n <= f(1) + F(n). If the integral converges, since F grows, it follows from the last inequalities s_n <= f(1) + F(n) <= f(1) + lim n tends to inf F(n) = = f(1) + $\displaystyle \int$ 1 to inf f(x) dx, which shows that the sequence s_n in bounded from above. Since it grows, it's convergent, so the inequalities for n->inf imply $\displaystyle \int$ 1 to inf f(x) dx <= $\displaystyle \sum f (n)$ <= f(1) + $\displaystyle \int$ 1 to inf f(x) dx. So, the series converges. So we get a set that has a point (I don't know the English term, I'll call it limiting point) such that for its any neighbourhood there's an element of it. (Any point s => 1). By a theorem from complex analysis, which states: Let f and g be analytic functions on a domain O (in this case it means an unempty open subset of C such that for any 2 points a, b in it there are finitely many points a = z0, z1, ..., zn = b such that [z0, z1], ...[zn-1, zn] lie in O) If f, g are equal on an infinite set which has a limiting point in O, then theyre equal on O, i.e. f =g. So this means) if we extend the series to C\{1} that whichever function equals it for s > 1, they're the same. Or, in other words, we can uniquely perform analytic continuation, if we can do it at all. (Which we can and is exectly what Riemann did first I think in the legendary paper you mentioned). And since every analytic function is continuous, by a theorem which states that any value of a continuous function equals its limit at that point, the zeros equal the limit of the function at them.
Just corrected an error.

 Tags hypothesis, questions, riemann, simple

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post robinlrandall New Users 2 August 23rd, 2014 03:10 PM mathbalarka Number Theory 0 October 31st, 2013 12:54 AM Eureka Number Theory 1 September 12th, 2011 01:07 PM MyNameIsVu Complex Analysis 2 April 16th, 2009 07:51 AM Whoever Number Theory 72 February 11th, 2009 07:03 AM

 Contact - Home - Forums - Top

Copyright © 2017 My Math Forum. All rights reserved.