Let Γ(s)=∫∞0ts−1e−tdt - classical gamma function - and γ(s,x)=∫x0ts−1e−tdt - lower incomplete gamma function. Then is the following true
limα→11Γ(1−α)γ(1−α,[Γ(1−α)]−1/α)=1
or false?
Intuitively, looking at the expression of γ and taking into account that limα→1Γ(1−α)=∞, I would say the limit is 0. Yet, if we can use the fact that γ(s,x) near x=0 behaves like xs/s, this would lead us to
limα→11Γ(1−α)γ(1−α,[Γ(1−α)]−1/α)=limα→1[Γ(1−α)]−1/α1−α=limα→1[Γ(α)sin(πα)/π]1/α1−α=limα→1[sin(πα)/π]1/α1−α.
Then, using the fact that when x is close to π, sin(x)≈π−x, I would go on with
=limα→1[(π−πα)/π]1/α1−α=limα→1(1−α)1/α1−α=limα→1exp{(1α−1)log(1−α)}
which eventually equals 1, since the exponent in the last line converges to 0. Am I loosing something/making some mistake?
Answer
Hint:
1x[Γ(x)]1/(1−x)=xx/(1−x)[xΓ(x)]1/(1−x)=xx/(1−x)[Γ(x+1)]1/(1−x)→01[Γ(1)]1=1
Since
limx→0xx/(1−x)=1
No comments:
Post a Comment