The core of the program consists of routines for adding a gluon to a
colour dipole, so as to produce two colour dipoles. The transverse
position and rapidity of the gluon are chosen randomly, in accord with
the distributions derived by Mueller in
Nucl. Phys. B415 (1994) 373.
The branching of one dipole into two is repeated recursively until the
rapidity of any new gluons would be larger than the rapidity to which
the onium is being evolved, giving the complete dipole structure of
the onium. This can be done for any number of initial onia, of arbitrary size.
The resulting dipole structures
can then be processed in whichever way the user wishes. For example
one can make a histogram of the number of dipoles as a function of
their size and position, to help understand which impact parameters
are relevant for onium-onium interaction.
One of the main advantages of using the Monte Carlo approach over
analytical techniques, is that it is very easy to determine quantities
which depend on the correlations between different dipoles in the same
onium. The OEDIPUS distribution includes example routines for doing
this.
In particular, there are highly optimised routines for determining the
multiple scattering interactions between pairs of onia. These store
the probability distribution of the single interaction (1-pomeron)
for each impact parameter allowing to reuse the data from a run to
determine many different quantities (e.g. two-pomeron interaction,
fully unitarised amplitude, cross section for any number of cut
pomerons, etc...).