now, synchronisation of joint departures done in two steps:
-first, plans optimised without taking care of synchronisation
-second, the penalty approach is applied, starting from the
optimum of first step.
This allows to use high unsynchronisation penalties without
getting traped