“…Next, we choose the samples at (t k , t l ) = (n 1,k ∆ 1 , n 2,l ∆ 2 ) = for k = 1, 2..., M 1 and l = 1, 2..., M 2 . To find the (t k , t l ) values, it is only necessary to compute the samples of the integrand in (12). Denoting the integrand by g(t 1 , t 2 ), we can express its samples as…”