 |  |
Astron. Astrophys. 331, 1103-1107 (1998)
4. Comparison with numerical results
Figs. 1a and 1b show how the theoretical results for
all agree with the numerical simulation based
on the method of MC. 10,000 realizations were used with 10,000 steps
in x. For both and
the agreement is so good that the theoretical
and numerical curves can hardly be distinguished. The very slight
differences arise for two reasons. Firstly, because of the finite
number of realisations there is a small residual noise component
present. Secondly, there is a cumulative error in the numerical
integration because of the finite step size in x.
![[FIGURE]](img76.gif) |
Fig. 1a and b. a and b are plotted for . In both graphs, the upper curve is for , the middle curve is for and the lower curve is for (for , is just zero).
|
© European Southern Observatory (ESO) 1998
Online publication: March 3, 1998
helpdesk.link@springer.de  |