raghav iyer

A cool problem

An integral representation of the Stieltjes constants:

We prove the identity

γn=1{x}x2(nlogx)logn1xdx.

We have

{x}=xifor ix<i+1.

So,

RHS&=i=1ii+1(xi)(nlogx)logn1xx2dx.

Now, since

(xi)(nlogx)logn1xx2dx=lognx(ixlogxn+1+1),

we obtain

ii+1(xi)(nlogx)logn1xx2dx=(i+1)logn+1i+((i+1)log(i+1)+n+1)logn(i+1)(i+1)(n+1).

Summing from i=1 to T, we get

RHS=limT((n(T+1)log(T+1)+1)logn(T+1)(n+1)(T+1)+i=2Tlognii).

Moreover,

logn(T+1)(n(T+1)log(T+1)+1)(n+1)(T+1)~logn+1Tn+1.

Therefore,

RHS=limT(i=1Tlogniilogn+1Tn+1).

By the defining limit for the n-th Stieltjes constant,

γn=1{x}x2(nlogx)logn1xdx.

Fuck you nassim.