I had asked this question sometime ago here. Now I have a question which I think is related to it.
Let f be an increasing function (continuous, of course!) with f(1)=0.
Consider the sequence sn=(n∑k=1f(k)−n∫1f(x)dx).
When does sn converge?
Answer
Qiaochu was on the right track to use an integral-to-sum formula, but it sounds like you want the Abel-Plana summation formula:
limn→∞(n∑k=mf(k)−∫nmf(u)du)=f(m)2−∫∞−∞(|t|exp(|2πt|)−1)(f(m+it)−f(m−it)2it)dt
This is used for instance to evaluate the Stieltjes constants. If the expression on the right hand side is convergent, then it is equivalent to the left hand side.
Adendum for Chandru:
Definitely f(z) should be analytic, or at least analytic in the region where ℜz≥m. Per Henrici's "Applied and Computational Complex Analysis", the additional conditions are
limt→∞f(u±it)=0
uniformly in u, and that
limt→∞|f(u±it)|exp(∓2πt)=0
uniformly in u.
EDIT:
For those scratching their head on just how Abel-Plana and Euler-Maclaurin are connected, the identity
∫∞−∞(|t|exp(|2πt|)−1)|t|2n−2dt=|B2n|2n
might be of interest.
No comments:
Post a Comment