164
I. P. Markov and M. V. Markina
Fig. 11.5 Amplitude function f 3
i j (ψ)
Fig. 11.6 Phase functions q m (ψ)
(Fig. 11.8), 8, 9, and 10 (Fig. 11.9). Errors err(I 1 ) and err(I 2 ) are very close to
each other for every order of EW method, and err(I 2 ) seems to be always slightly
bigger than err(I 1 ). As the total number of integrand evaluations increases, differences between errors from different orders of EW method are getting smaller and
smaller, suggesting that the proposed composite integration procedure converges in
number of subintervals and in order of underlining EW method. We point out that
already for total number of integrand evaluations ~400 both errors for all cases are
less than 10
−10 .