13.11.2014 Views

proofs

proofs

proofs

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

44 Three times π 2 /6<br />

v<br />

1<br />

1<br />

2<br />

y<br />

v<br />

1<br />

u<br />

x<br />

This evaluation also shows that the double integral (over a positive function<br />

with a pole at x = y = 1) is finite. Note that the computation is also easy<br />

and straightforward if we read it backwards — thus the evaluation of ζ(2)<br />

leads one to the double integral I.<br />

The second way to evaluate I comes from a change of coordinates: in the<br />

new coordinates given by u := y+x<br />

2<br />

and v := y−x<br />

2<br />

the domain of integration<br />

is a square of side length 2√ 1 2, which we get from the old domain by<br />

first rotating it by 45 ◦ and then shrinking it by a factor of √ 2. Substitution<br />

of x = u − v and y = u + v yields<br />

1<br />

1 − xy = 1<br />

1 − u 2 + v 2 .<br />

To transform the integral, we have to replace dxdy by 2 du dv, to compensate<br />

for the fact that our coordinate transformation reduces areas by a<br />

constant factor of 2 (which is the Jacobi determinant of the transformation;<br />

see the box on the next page). The new domain of integration, and the<br />

function to be integrated, are symmetric with respect to the u-axis, so we<br />

just need to compute two times (another factor of 2 arises here!) the integral<br />

over the upper half domain, which we split into two parts in the most<br />

natural way:<br />

I 1<br />

1<br />

2<br />

I 2<br />

1<br />

u<br />

Using<br />

∫<br />

I = 4<br />

1/2( u<br />

0<br />

∫<br />

0<br />

) ∫<br />

dv<br />

1<br />

1 − u 2 + v 2 du + 4<br />

1/2<br />

( 1−u ∫<br />

0<br />

)<br />

dv<br />

1 − u 2 + v 2 du.<br />

∫<br />

dx<br />

a 2 + x 2 = 1 a arctan x + C , this becomes<br />

a<br />

∫1/2<br />

( )<br />

1<br />

I = 4 √ arctan u<br />

√ du<br />

1 − u<br />

2 1 − u<br />

2<br />

+ 4<br />

0<br />

∫ 1<br />

( )<br />

1 1 − u<br />

√ arctan √ du.<br />

1 − u<br />

2 1 − u<br />

2<br />

1/2<br />

These integrals can be simplified and finally evaluated by substituting u =<br />

sin θ resp. u = cosθ. But we proceed more directly, by computing that the<br />

derivative of g(u) := arctan ( )<br />

√ u<br />

1−u 2 is g ′ (u) = 1 , while the deriva-<br />

√<br />

1−u 2<br />

tive of h(u) := arctan ( (<br />

√1−u<br />

1−u 2) √ = arctan<br />

1−u<br />

1+u)<br />

is h ′ (u) = − 1 √<br />

2<br />

. 1−u 2<br />

So we may use ∫ b<br />

a f ′ (x)f(x)dx = [ 1<br />

2 f(x)2] b<br />

a = 1 2 f(b)2 − 1 2 f(a)2 and get<br />

I = 4<br />

∫ 1/2<br />

0<br />

= 2<br />

[g(u) 2] 1/2<br />

g ′ (u)g(u)du + 4<br />

0<br />

∫ 1<br />

1/2<br />

− 4<br />

[h(u) 2] 1<br />

1/2<br />

−2h ′ (u)h(u)du<br />

1<br />

= 2g( 1 2 )2 − 2g(0) 2 − 4h(1) 2 + 4h( 1 2 )2<br />

= 2 ( )<br />

π 2 (<br />

6 − 0 − 0 + 4 π<br />

) 2<br />

6<br />

= π2<br />

6 . □

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!