Could you or someone else tell me if I'm interpreting this correctly since I've never worked on one like this before. But first, let's just restrict it to a double integral for now:
\mathop\iint\limits_{D} f(k,q)d^3q\, d^3 k
where each integral is a triple integral in spherical coordinates and:
D={|k|>1, |k+q|<1}
We can compute the outer one easily. Since |k|>1, then for spherical coordinate r, we can write:
\mathop\int_{r>1}\left( \mathop\int\limits_{S} f(k,q) d^3 q\right)\,d^3 k
So what is S? Since |k+q|<1, then that means we need:
\sqrt{(k_x+q_z)^2+(k_y+q_y)^2+(k_z+q_z)^2}<1
for every point in k-space (k_x, k_y, k_z). Now suppose we have for a particular point:
k=(3,4,7)
Then for |k+q|<1, we would have to integrate in q-space over a sphere centered at q=(-3,-4,-7) with radius one. The boundary for that one k-point would be:
(q_x+3)^2+(q_y+4)^2+(q_z+7)^2=1
So for just that one k-point, the integral would be:
\mathop\iiint\limits_{(q_x+3)^2+(q_y+4)^2+(q_z+7)^2\leq 1} f(k,q)d^3q
and therefore for all of the k-space, we could then write:
\mathop\iiint\limits_{k_x^2+k_y^2+k_z^2>1}\iiint_{(q_x-k_x)^2+(q_y-k_y)^2+(q_z-k_z)^2\leq 1} f(k,q)d^3q d^3k
Ok, so just for now, can we let p be what ever it has to be to work, say p=(1,1,1) or whatever, can we now compute:
\mathop\iiint\limits_{k_x^2+k_y^2+k_z^2>1}\iiint_{(q_x-k_x)^2+(q_y-k_y)^2+(q_z-k_z)^2\leq 1} \exp\{-(q^2+q\cdot k+q\cdot(1,1,1))\}d^3q d^3k