In his gorgeous paper "How to compute ∑1n2 by solving triangles", Mikael Passare offers this idea for proving ∑∞n=11n2=π26:
Proof of equality of square and curved areas is based on another picture:
Recapitulation of Passare's proof using formulas is as follows:
∞∑n=11n2=∞∑n=1∫∞0e−nxndx=−∫∞0log(1−e−x)dx=π26
There is also another paper dealing with geometric proof of ∑∞n=11n2=π26, in an entirely different way.
I tried to find a similar way to prove:
∞∑n=11n4=π490
but didn't succeed. Maybe you will?
Answer
The first part is similar.
1n4=1n∫∞0∫∞0∫∞0e−n(x+y+z)dxdydz
so
∞∑n=11n4=∫∞0∫∞0∫∞0−log(1−e−(x+y+z))dxdydz
Now we're integrating over an octant of R3. Change variables to u=x, v=x+y, w=x+y+z, with dudvdw=dxdydz:
∞∑n=11n4=∫∞w=0∫wv=0∫vu=0−log(1−e−w)dudvdw=−∫∞0w2log(1−e−w)2dw
The tricky part is evaluating that integral.
No comments:
Post a Comment