| Author |
Comment/Response |
Sam
|
 01/19/11 7:15pm
I'm doing an integral that I believe should just evaluate to a real number, and am getting a strange pi*i term in the result. The integrand is a function in [0,1], and the boundaries of integration are all within [0,1]. Here is the code I am running:
abstract =
Piecewise[{{0, a >= d && a <= b}, {0, b >= e && b <= a}, {0,
d >= a && d <= e}, {0, e >= b && e <= d}},
c*(e - d)/(a - b - d + e) + f*(a - b)/(a - b - d + e)]
Integrate[abstract, {c, 0, 1}, {d, 0, 1}, {e, 0, 1}, {a, c, 1}, {b, c,
1}, {f, 0, c}, {j, 0, c}]
The output simplifies to -(2869/57600) + (111 I \[Pi])/1280 + (8 Log[2])/105
Interestingly, the result agrees with my simulation if I ignore the 111 pi*i / 1280 term.
Any idea why this pi*i term is showing up? Thanks!
URL: , |
|
