Positions of a vessels are calculated using a numerical integration method called runge-kutta. All major g-bodys including the moons of the jupiter are calculated using ephemerides. Ephemeris is not a numerical integration but it do have a similar precission/accuracy than a numerical integration have. Some other minor bodies those aren't included in the ephemeris package are approximated as a 2-body system.
The moons of the jupiter do behave like being attracted by other g-bodies. But if you add a new moon into the jupiter system that doesn't exist in reality it won't attract the other moons.
Calculations made by TransX are based on a 2-body techniques. They are only a rough approximations of the vessel's trajectory. Errors will increase as the time of prediction is increasing. 2-body calculatios are accurate only when the vessel is deep within the gravitational field of a g-body.
Time acceleration is not effecting in a calculations. There aren't any other things causing a significant effects.
Yes, there is. You can use Orbiter's Celbody interface but it will only work for g-bodies those are controlled by ephemerides.
Like this:
CELBODY *cMoon = oapiGetCelbodyInterface(hMoon);
cMoon->clbkEphemeris(MJD, EPHEM_TRUEPOS, &pos_vel_data);
Look more details from the SDK documentation.
