PDA

View Full Version : Problems with Riemann function...


eljose79
Aug7-04, 09:32 AM
let be the product R(s)R(s+a) with a a complex or real number..the i would like to know the limit Lim(s tends to e) being e a number so R(e)=0 żis there a number a so the limit is non-zero nor infinite?..thanks.

shmoe
Aug7-04, 09:44 AM
It will only potentially be non-zero when \zeta(s) has a pole at e+a. Of course this only happens when e+a=1. This limit will never be infinite, since this lone pole of \zeta(s) is simple.

I say "potentially" non-zero, because the zero could have order 2 or more, making the limit zero. All known zeros to date have been simple, and it's suspected that they all are.

eljose79
Aug8-04, 06:33 AM
Another question..could be proved that 1/R(s+1/2)=O(R(s)?

shmoe
Aug8-04, 07:55 AM
Another question..could be proved that 1/R(s+1/2)=O(R(s)?

You mean \displaystyle\frac{1}{\zeta(1/2+it)}=O(\zeta(1/2+it))? If so, no, since the left side has infinitely many poles.

Do you mean something else? I'm not sure what values of s you're considering, and the +1/2 makes me think of the critical line, hence my guess to your meaning.

eljose79
Aug9-04, 04:59 AM
thanks again.... the equality 1/R(1/2+it)=O(R(1/2+it) is true?..

another question let be the Riemann zeta function inside the criticla strip 0<sigma<1 then ..is the product R(a+s)R(s) bounded in the sense exist a and b so a<[R(s+a)R(s)]<b where [x] is the modulus of x [x]=sqrt(x*.x) ?

shmoe
Aug9-04, 09:18 AM
thanks again.... the equality 1/R(1/2+it)=O(R(1/2+it) is true?..

Absolutely not. 1/\zeta(1/2+it) has poles for infinitely many values of t.


another question let be the Riemann zeta function inside the criticla strip 0<sigma<1 then ..is the product R(a+s)R(s) bounded in the sense exist a and b so a<[R(s+a)R(s)]<b where [x] is the modulus of x [x]=sqrt(x*.x) ?

You're going to have to clarify what you mean by sqrt(x*.x).

There are bounds of the form \zeta(\sigma+it)=O(|t|^{k}) for any \sigma fixed.The exponent k depends on \sigma. For example, if \sigma=1/2, then k=1/6 will work (something slightly smaller than 1/6 is known to be true). If \sigma >1 then k=0 will work.