214k views
20 votes
How do you solve this?

How do you solve this?-example-1
User Davidmpaz
by
3.9k points

1 Answer

4 votes

Answer:

Step 1. If

I1(n)=∑1≤k≤n√(Γ(kn))−k

Then limn→∞I1(n)=0.

Proof. Indeed, since Γ is decreasing on (0,1] we have

I1(n)≤∑1≤k≤n√(Γ(1n−−√))−k≤∑k=1∞(Γ(1n−−√))−k=1Γ(1/n−−√)−1

and step 1. follows.

Step 2. If

I2(n)=∑n√<k≤n/2(Γ(kn))−k

Then limn→∞I2(n)=0.

Proof. Recall that Γ attains its minimum ≈0.8856, on [1,2], at some some point x0≈1.4616. In particular, Γ(x)≥2/3 for 1≤x≤2. So, for n−−√<k≤n/2 we have

knΓ(kn)=Γ(1+kn)≥23

Thus, for n−−√<k≤n/2, we have Γ(k/n)>4/3. It follows that

I2(n)≤∑k>n√(34)k=4(34)⌈n√⌉

and step 2. follows.

Step 3. If

I3(n)=∑n/2<k≤n(Γ(kn))−k

Then limn→∞I3(n)=eγeγ−1. where γ is the Euler-Mascheroni constant.

Proof. Note first that, with p=n−k,

I3(n)=∑0≤p<n/2(Γ(1−pn))p−n=∑p=0∞ap(n)

with

ap(n)={(Γ(1−pn))p−n0ifotherwise0≤p<n/2

Now, since Γ(1)=1 and Γ′(1)=−γ we have, for a fixed p and large n:

(p−n)lnΓ(1−pn)=(p−n)ln(1+γpn+O(1n2))=−γp+O(1n)

Thus

∀p≥0,limn→∞ap(n)=e−γp.(1)

desired conclusion follows:

limn→∞∑1≤k≤n(Γ(kn))−k=limn→∞(I1(n)+I2(n)+I3(n))=eγeγ−1.

Explanation:

User Derek Hill
by
4.2k points