Processing math: 40%

Pólya-Vinogradov inequality


Theorem 1.

For m,nN and p a positive odd rational prime,

|m+nt=m(tp)|<plnp.
Proof.

Start with the following manipulations:

m+nt=m(tp)=1pp-1t=0m+nx=mp-1a=0(tp)e2πia(x-t)/p=1pp-1a=1m+nx=me2πiax/pp-1t=0(tp)e-2πiat/p

The expression p-1t=0(tp)e-2πiat/p is just a Gauss sumDlmfPlanetmath, and has magnitude p. Hence

|m+nt=m(tp)| |ppp-1a=1m+nx=me2πaix/p|=|ppp-1a=1e2πiam/pnx=0e2πiax/p||ppp-1a=1e2πian/p-1e2πia/p-1|
= |ppp-1a=1eπian/psin(πan/p)eπia/psin(πa/p)|ppp-1a=1|1sin(πa/p)|ppp-1a=112a/p

Here x denotes the absolute valueMathworldPlanetmathPlanetmathPlanetmath of the difference between x and the closest integer to x, i.e. x=inf.

Since p is odd, we have

12a=1p-11a/p=0<a<p2pa=pa=1p-121a

Now ln2x+12x-1>1x for x>1; to prove this, it suffices to show that the functionMathworldPlanetmath f:[1,)𝐑 given by f(x)=xln2x+12x-1 is decreasing and approaches 1 as x. To prove the latter statement, substitute v=1/x and take the limit as v0 using L’Hôpital’s rule. To prove the former statement, it will suffice to show that f is less than zero on the interval [1,). But f(x)0 as x and f is increasing on [1,), since f′′(x)=-44x2-1(1-4x2+14x2-1)>0 for x>1, so f is less than zero for x>1.

With this in hand, we have

|t=mm+n(tp)|pppa=1p-121a<pa=1p-12ln2a+12a-1=plnp.

References

Title Pólya-Vinogradov inequality
Canonical name PolyaVinogradovInequality
Date of creation 2013-03-22 12:46:23
Last modified on 2013-03-22 12:46:23
Owner djao (24)
Last modified by djao (24)
Numerical id 7
Author djao (24)
Entry type Theorem
Classification msc 11L40