Let the sequence u1,u2,ldots satisfy un+1=un−u2n+O(u3n). Then it can be shown that if unto0 as ntoinfty, then un=n−1+O(n−2logn). (See N. G. de Bruijn, Asymptotic methods in analysis, Section 8.5.)
This can be used to obtain asymptotics for vn+1=Avn−Bv2n+O(v3n), where A and B are constants. Let wn=A−nvn; this gives
An+1wn+1=An+1wn−BAnw2n+O(Anw3n)
and so
wn+1=wn−BA−1w2n+O(w3n).
Then let wn=Axn/B to get
Axn+1/B=Axn/B−B/Acdot(Axn/B)2+O(x3n)
and after simplifying
xn+1=xn−x2n+O(x3n). This satisfies the initial requirements for un (with some checking of the side condition); then substitute back.
But say I actually know that un+1=P(un) for some polynomial P, with P(z)=z−z2+a3z3+cdots+adzd. In this case it seems like it should be possible to get more explicit information about un. Is there a known algorithm for computing an asymptotic series for un as ntoinfty?
No comments:
Post a Comment